feat(browser-extension): add opencode-browser-cli npm helper

`opencode browser install` only exists on this branch, so nobody on a released opencode can connect the extension. packages/browser-extension/cli is a standalone npm package (opencode-browser-cli) that does the same setup without changing opencode: - install: copies itself to ~/.local/share/opencode-browser (npx caches can be cleared), registers the ai.opencode.browser host for every installed Chromium browser (macOS/Linux manifests, Windows registry), adds the Browser Control MCP to opencode's global config (jsonc edits keep comments), starts the service, and copies the bundled unpacked extension to a permanent folder for Load unpacked. - host: answers the extension with the URL and password from the user's installed `opencode service start` / `service get password`, and installs the extension's plugin. - status, extension (opens the folder), uninstall. The bundle includes jsonc-parser and the extension build (manifest key kept so the unpacked ID matches the helper). The panel, welcome tab, store copy, and privacy policy now point at `npx opencode-browser-cli install`. The opencode CLI command is unchanged. Verified: lint, extension typecheck; `npm pack` then npx from the tarball in a temp HOME/XDG registered Helium, kept a commented opencode.json and added the MCP entry, started the service; the wrapper answered service (URL + password) and plugin (written) messages over native messaging; status and uninstall worked.
chore(browser-extension): capitalize OpenCode in the store summary
2026-10-03 14:06:29 +00:00 · 2026-10-02 20:38:40 -04:00 · 2026-10-02 14:37:45 -04:00 · 2026-10-02 14:34:52 -04:00 · 2026-10-02 14:33:28 -04:00 · 2026-10-02 14:13:35 -04:00
2318 changed files with 103407 additions and 96095 deletions
@@ -1,5 +0,0 @@
---
-"@opencode/core": patch
---
-
-Correct directory page headings when the read offset is zero.
@@ -45,14 +45,16 @@ runs:
    - name: Get cache directory
      id: cache
      shell: bash
-      run: echo "dir=$(bun pm cache)" >> "$GITHUB_OUTPUT"
+      run: |
+        echo "dir=$(bun pm cache)" >> "$GITHUB_OUTPUT"
+        echo "version=$(bun --version)" >> "$GITHUB_OUTPUT"

    - name: Restore Bun dependencies
      id: bun-cache
      uses: actions/cache/restore@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0
      with:
        path: ${{ steps.cache.outputs.dir }}
-        key: ${{ runner.os }}-bun-${{ hashFiles('**/bun.lock') }}
+        key: ${{ runner.os }}-${{ runner.arch }}-bun-${{ steps.cache.outputs.version }}-${{ hashFiles('bun.lock', 'patches/**') }}

    - name: Install setuptools for distutils compatibility
      run: python3 -m pip install setuptools || pip install setuptools || true
@@ -75,4 +77,4 @@ runs:
      uses: actions/cache/save@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0
      with:
        path: ${{ steps.cache.outputs.dir }}
-        key: ${{ runner.os }}-bun-${{ hashFiles('**/bun.lock') }}
+        key: ${{ steps.bun-cache.outputs.cache-primary-key }}
@@ -7,6 +7,10 @@ on:
    branches: [dev, v2]
  workflow_dispatch:

+concurrency:
+  group: ${{ case(github.ref == 'refs/heads/dev', format('{0}-{1}', github.workflow, github.run_id), format('{0}-{1}', github.workflow, github.event.pull_request.number || github.ref)) }}
+  cancel-in-progress: true
+
 jobs:
  check:
    name: typecheck
@@ -15,7 +15,7 @@ jobs:
  close-non-compliant:
    runs-on: ubuntu-latest
    steps:
-      - name: Close non-compliant issues and PRs after 2 hours
+      - name: Close non-compliant issues and PRs after 72 hours
        uses: actions/github-script@f28e40c7f34bde8b3046d885e986cb6290c5673b # v7.1.0
        with:
          script: |
@@ -33,7 +33,7 @@ jobs:
            }

            const now = Date.now();
-            const twoHours = 2 * 60 * 60 * 1000;
+            const seventyTwoHours = 72 * 60 * 60 * 1000;
            const orgMemberAssociations = new Set(['OWNER', 'MEMBER']);
            const agentLogin = 'opencode-agent[bot]';
            const { data: file } = await github.rest.repos.getContent({
@@ -87,14 +87,14 @@ jobs:
              if (!complianceComment) continue;

              const commentAge = now - new Date(complianceComment.created_at).getTime();
-              if (commentAge < twoHours) {
-                core.info(`${kind} #${item.number} still within 2-hour window (${Math.round(commentAge / 60000)}m elapsed)`);
+              if (commentAge < seventyTwoHours) {
+                core.info(`${kind} #${item.number} still within 72-hour window (${Math.round(commentAge / 60000)}m elapsed)`);
                continue;
              }

              const closeMessage = isPR
-                ? 'This pull request has been automatically closed because it was not updated to meet our [contributing guidelines](../blob/dev/CONTRIBUTING.md) within the 2-hour window.\n\nFeel free to open a new pull request that follows our guidelines.'
-                : 'This issue has been automatically closed because it was not updated to meet our [contributing guidelines](../blob/dev/CONTRIBUTING.md) within the 2-hour window.\n\nFeel free to open a new issue that follows our issue templates.';
+                ? 'This pull request has been automatically closed because it was not updated to meet our [contributing guidelines](../blob/dev/CONTRIBUTING.md) within the 72-hour window.\n\nFeel free to open a new pull request that follows our guidelines.'
+                : 'This issue has been automatically closed because it was not updated to meet our [contributing guidelines](../blob/dev/CONTRIBUTING.md) within the 72-hour window.\n\nFeel free to open a new issue that follows our issue templates.';

              await github.rest.issues.createComment({
                owner: context.repo.owner,
@@ -129,5 +129,5 @@ jobs:
                });
              }

-              core.info(`Closed non-compliant ${kind} #${item.number} after 2-hour window`);
+              core.info(`Closed non-compliant ${kind} #${item.number} after 72-hour window`);
            }
@@ -107,7 +107,7 @@ jobs:

          If the issue is NOT compliant and the author association is not OWNER or MEMBER, start the comment with:
          <!-- issue-compliance -->
-          Then explain what needs to be fixed and that they have 2 hours to edit the issue before it is automatically closed. Also add the label needs:compliance to the issue using: gh issue edit ${{ github.event.issue.number }} --add-label needs:compliance
+          Then explain what needs to be fixed and that they have 72 hours to edit the issue before it is automatically closed. Also add the label needs:compliance to the issue using: gh issue edit ${{ github.event.issue.number }} --add-label needs:compliance

          If duplicates were found, include a section about potential duplicates with links.

@@ -124,7 +124,7 @@ jobs:
          **What needs to be fixed:**
          - [specific reasons]

-          Please edit this issue to address the above within **2 hours**, or it will be automatically closed.
+          Please edit this issue to address the above within **72 hours**, or it will be automatically closed.

          [If duplicates found, add:]
          ---
@@ -37,7 +37,7 @@ jobs:
          echo "=== Flake structure ==="
          nix flake show --all-systems

-          SYSTEMS="x86_64-linux aarch64-linux x86_64-darwin aarch64-darwin"
+          SYSTEMS="x86_64-linux aarch64-linux aarch64-darwin"
          PACKAGES="opencode"
          # TODO: move 'desktop' to PACKAGES when #11755 is fixed
          OPTIONAL_PACKAGES="desktop"
@@ -34,8 +34,6 @@ jobs:
            runner: blacksmith-4vcpu-ubuntu-2404
          - system: aarch64-linux
            runner: blacksmith-4vcpu-ubuntu-2404-arm
-          - system: x86_64-darwin
-            runner: macos-15-intel
          - system: aarch64-darwin
            runner: macos-latest
    runs-on: ${{ matrix.runner }}
@@ -126,7 +124,7 @@ jobs:

          [ -f "$HASH_FILE" ] || echo '{"nodeModules":{}}' > "$HASH_FILE"

-          for SYSTEM in x86_64-linux aarch64-linux x86_64-darwin aarch64-darwin; do
+          for SYSTEM in x86_64-linux aarch64-linux aarch64-darwin; do
            FILE="hashes/hash-${SYSTEM}/hash.txt"
            if [ -f "$FILE" ]; then
              HASH="$(tr -d '[:space:]' < "$FILE")"
@@ -307,7 +307,7 @@ jobs:
            **What needs to be fixed:**
            ${issues.map(i => `- ${i}`).join('\n')}

-            Please edit this PR description to address the above within **2 hours**, or it will be automatically closed.
+            Please edit this PR description to address the above within **72 hours**, or it will be automatically closed.

            If you believe this was flagged incorrectly, please let a maintainer know.`;

@@ -24,6 +24,10 @@ on:
        description: "Override version (optional)"
        required: false
        type: string
+      release_notes:
+        description: "Reviewed V2 release notes for the Discord announcement (optional)"
+        required: false
+        type: string

 concurrency: ${{ github.workflow }}-${{ github.ref }}-${{ (github.ref_name == 'v2' && (inputs.version || inputs.bump) && 'release') || inputs.version || inputs.bump }}

@@ -653,3 +657,19 @@ jobs:
          OPENCODE_DESKTOP_DIST: /tmp/desktop
          CLOUDFLARE_ACCOUNT_ID: 15d29c8639fd3733b1b5486a2acfd968
          CLOUDFLARE_API_TOKEN: ${{ secrets.CLOUDFLARE_API_TOKEN }}
+
+  notify-discord-v2:
+    needs:
+      - version
+      - publish
+    if: ${{ !cancelled() && github.repository == 'anomalyco/opencode' && github.ref_name == 'v2' && needs.version.outputs.release && needs.publish.result == 'success' }}
+    runs-on: blacksmith-4vcpu-ubuntu-2404
+    steps:
+      # Unlike dev, V2 publishes a tag rather than a GitHub Release event.
+      - name: Announce V2 release in Discord
+        uses: SethCohen/github-releases-to-discord@24d166886aee4646d448c8a389ff9e1ebcab3682 # v1.20.0
+        with:
+          webhook_url: ${{ secrets.DISCORD_WEBHOOK }}
+          release_name: OpenCode V2 ${{ needs.version.outputs.tag }}
+          release_body: ${{ inputs.release_notes }}
+          release_html_url: https://github.com/${{ github.repository }}/tree/${{ needs.version.outputs.tag }}
@@ -94,12 +94,6 @@ jobs:
          git config --global user.email "bot@opencode.ai"
          git config --global user.name "opencode"

-      - name: Install ffmpeg
-        if: runner.os == 'Linux'
-        run: |
-          sudo apt-get update
-          sudo apt-get install --yes ffmpeg
-
      - name: Cache Turbo
        uses: actions/cache@0057852bfaa89a56745cba8c7296529d2fc39830 # v4.3.0
        with:
@@ -251,6 +245,13 @@ jobs:
        run: bunx playwright test --config e2e/service-worker/playwright.config.ts
        timeout-minutes: 5

+      - name: Run session UI component tests
+        if: ${{ !cancelled() && env.E2E_ENABLED == 'true' }}
+        run: bun --cwd packages/session-ui test:components
+        env:
+          CI: true
+        timeout-minutes: 15
+
      - name: Upload Playwright artifacts
        if: always() && env.E2E_ENABLED == 'true'
        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
@@ -261,3 +262,5 @@ jobs:
          path: |
            packages/app/e2e/test-results
            packages/app/e2e/playwright-report
+            packages/session-ui/component-tests/test-results
+            packages/session-ui/component-tests/playwright-report
@@ -1,5 +1,9 @@
 {
  "$schema": "https://raw.githubusercontent.com/nicolo-ribaudo/oxc-project.github.io/refs/heads/json-schema/src/public/.oxlintrc.schema.json",
+  "jsPlugins": [
+    { "name": "anti-slop", "specifier": "./script/oxlint/anti-slop/index.ts" },
+    { "name": "anti-slop-effect", "specifier": "./script/oxlint/anti-slop/effect/index.ts" }
+  ],
  "categories": {
    "correctness": "off",
    "suspicious": "off",
@@ -18,5 +22,94 @@
      }
    ]
  },
-  "ignorePatterns": ["**/node_modules", "**/dist", "**/.build", "**/.sst", "**/*.d.ts", "**/sdk.gen.ts"]
+  "overrides": [
+    {
+      "files": [
+        "packages/app/**",
+        "packages/desktop/**",
+        "packages/gui-extensions/**",
+        "packages/ui/**",
+        "packages/session-ui/**"
+      ],
+      "rules": {
+        "oxc/no-accumulating-spread": "warn",
+        "anti-slop/no-array-filter-map": "warn",
+        "anti-slop/no-reduce-accumulator-copy": "warn",
+        "anti-slop/no-chained-type-assertions": "warn",
+        "anti-slop/no-conditional-empty-object-spread": "warn",
+        "anti-slop/no-known-value-widening": "warn",
+        "anti-slop/no-module-mocking": "warn",
+        "anti-slop/no-object-parameters": "warn",
+        "anti-slop/no-reflect-apply": "warn",
+        "anti-slop/no-reflect-get": "warn",
+        "anti-slop/no-runtime-typeof": "warn",
+        "anti-slop/no-shape-in-symbol-names": "warn",
+        "anti-slop/no-unknown-parameters": "warn",
+        "anti-slop/no-unknown-returns": "warn",
+        "anti-slop/no-unknown-type-aliases": "warn",
+        "anti-slop/no-unsafe-dictionary-type": "warn",
+        "anti-slop/no-widen-then-assert": "warn",
+        "anti-slop/require-readable-spacing": "warn",
+        "anti-slop/require-safety-comment-for-type-assertion": "warn"
+      }
+    },
+    {
+      "files": ["packages/app/**", "packages/desktop/**", "packages/gui-extensions/**", "packages/session-ui/**"],
+      "rules": {
+        "anti-slop-effect/no-manual-effect-error-tag": "warn",
+        "anti-slop-effect/no-manual-tag-comparison": "warn",
+        "anti-slop-effect/no-manual-tagged-construction": "warn",
+        "anti-slop-effect/no-service-constructor-imports": "warn",
+        "anti-slop-effect/prefer-effect-match": "warn"
+      }
+    },
+    {
+      "files": ["packages/gui-extensions/src/*/**"],
+      "rules": {
+        "no-restricted-imports": [
+          "error",
+          {
+            "patterns": [
+              {
+                "regex": "^@opencode/(app|desktop)(/|$)",
+                "message": "GUI extensions never import the app or desktop packages. Use the SDK."
+              },
+              {
+                "regex": "^@/",
+                "message": "GUI extensions never import app internals. Use the SDK."
+              },
+              {
+                "group": ["../*/*", "!../*/contract", "!../sdk/*"],
+                "message": "Import another extension only through its contract.ts."
+              },
+              {
+                "regex": "\\.css$",
+                "message": "Import CSS with ?inline and contribute it with ctx.add(Style, css)."
+              }
+            ]
+          }
+        ]
+      }
+    }
+  ],
+  "ignorePatterns": [
+    "**/node_modules",
+    "**/dist",
+    "**/.build",
+    "**/.sst",
+    "**/*.d.ts",
+    "**/sdk.gen.ts",
+    ".agent/**",
+    ".agents/**",
+    ".claude/**",
+    ".codex/**",
+    ".continue/**",
+    ".cursor/**",
+    ".gemini/**",
+    ".opencode/**",
+    ".pi/**",
+    ".roo/**",
+    ".windsurf/**",
+    "script/oxlint/anti-slop/**"
+  ]
 }
@@ -184,7 +184,7 @@ const table = sqliteTable("session", {
 - Keep `SessionRunner`, model resolution, tool registry, permissions, and filesystem Location-scoped. Omitted `Location.workspaceID` means implicit-local placement; explicit workspace identity remains reserved for future placement semantics.
 - Preserve one explicit `llm.stream(request)` call per Physical Attempt and reload projected history before durable continuation. A logical Step may use generic pre-output retries, one full-context retry after continuation rejection, incomplete-stream continuation, or one overflow-compaction rebuild. Generic retries retain the logical step number and do not consume another agent-step allowance. Do not delegate orchestration to an in-memory tool loop.
 - Keep local Session drains process-local until clustering is implemented. `SessionRunCoordinator` joins explicit same-Session resumes, coalesces prompt wakeups, and allows different Sessions to run concurrently. A write-ahead execution claim marks a process-local busy period for restart recovery: terminal completion, failure, or user interruption releases it, while shutdown interruption and process death preserve it. Startup recovery resumes claimed top-level Sessions with durable per-execution attempt accounting. The claim is a recovery marker, not clustered ownership, fencing, or an exactly-once guarantee.
- Keep native compaction mechanisms out of `SessionCompaction`. Plugins register `native` strategies through the `SessionCompaction` editor that turn a prepared request into a replacement window (the built-in `NativeCompactionPlugin` handles `@opencode/ai` compaction operations); later registrations win. Core owns the provider-mode decision, route provenance, the retry policy, overflow recovery, interruption, usage accounting, and checkpoint persistence.
+- Keep provider-specific native compaction mechanisms in `@opencode/ai` behind `LLMClient.compact`. `SessionCompaction` chooses a summary or native compaction from the model's `compaction` setting and owns route provenance, request shrinking, the retry policy, interruption, usage accounting, and checkpoint persistence.
 - Keep delivery vocabulary explicit. Prompts steer by default. At safe step boundaries, steered compaction takes priority up to the first steered move control; other steers retain enqueue order. At an idle boundary, steers take priority; otherwise exactly one queued item delivers before the runner reevaluates continuation. Inbox items may be cancelled or changed between queue and steer before delivery. Promoting new user input resets the selected agent's step allowance; a batch of steers resets it once.
 - One step is one logical LLM call; its durable record covers only the model-visible span. Do not write "provider turn", and do not use bare "turn" for a single call: "turn" is reserved for the future assistant-turn unit containing all steps from prompt promotion until the session would go idle.
 - Keep event replay ownership separate from clustered Session execution ownership.
@@ -15,6 +15,7 @@
        "@actions/artifact": "5.0.1",
        "@ast-grep/cli": "0.44.0",
        "@opencode/client": "workspace:*",
+        "@oxlint/plugins": "1.60.0",
        "@tsconfig/bun": "catalog:",
        "@types/mime-types": "3.0.1",
        "@types/react": "19.2.17",
@@ -32,7 +33,7 @@
    },
    "packages/ai": {
      "name": "@opencode/ai",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@aws-sdk/credential-providers": "3.1057.0",
        "@opencode/schema": "workspace:*",
@@ -54,7 +55,7 @@
    },
    "packages/app": {
      "name": "@opencode/app",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@corvu/drawer": "catalog:",
        "@dnd-kit/abstract": "0.5.0",
@@ -64,7 +65,7 @@
        "@ibm/plex": "6.4.1",
        "@kobalte/core": "catalog:",
        "@opencode/client": "workspace:*",
-        "@opencode/plugin-browser": "workspace:*",
+        "@opencode/gui-extensions": "workspace:*",
        "@opencode/schema": "workspace:*",
        "@opencode/session-ui": "workspace:*",
        "@opencode/ui": "workspace:*",
@@ -86,13 +87,11 @@
        "core-js": "3.50.0",
        "effect": "catalog:",
        "fuzzysort": "catalog:",
-        "ghostty-web": "github:anomalyco/ghostty-web#83c0a07b8628b748aed073b232cb4b52a6ca11c1",
        "qr-scanner": "1.4.2",
        "remeda": "catalog:",
        "solid-js": "catalog:",
        "solid-presence": "0.2.0",
        "tailwindcss": "4.3.3",
-        "uqr": "0.1.3",
      },
      "devDependencies": {
        "@happy-dom/global-registrator": "20.0.11",
@@ -103,6 +102,7 @@
        "@types/node": "catalog:",
        "@typescript/native-preview": "catalog:",
        "diff": "catalog:",
+        "ghostty-web": "github:anomalyco/ghostty-web#83c0a07b8628b748aed073b232cb4b52a6ca11c1",
        "happy-dom": "20.11.1",
        "tw-animate-css": "1.4.0",
        "vite": "8.2.2",
@@ -110,15 +110,41 @@
        "vite-plugin-solid": "2.11.14",
      },
    },
+    "packages/browser-extension": {
+      "name": "@opencode/browser-extension",
+      "version": "2.0.22",
+      "dependencies": {
+        "@opencode/client": "workspace:*",
+        "@opencode/plugin-browser": "workspace:*",
+        "@opencode/session-ui": "workspace:*",
+        "@opencode/ui": "workspace:*",
+        "effect": "catalog:",
+        "solid-js": "catalog:",
+      },
+      "devDependencies": {
+        "@opencode/plugin": "workspace:*",
+        "@tailwindcss/vite": "4.3.3",
+        "@tsconfig/node22": "catalog:",
+        "@types/bun": "catalog:",
+        "@types/chrome": "0.3.4",
+        "@typescript/native-preview": "catalog:",
+        "devtools-protocol": "0.0.1687809",
+        "tailwindcss": "catalog:",
+        "typescript": "catalog:",
+        "vite": "8.2.2",
+        "vite-plugin-solid": "2.11.14",
+      },
+    },
    "packages/cli": {
      "name": "@opencode/cli",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "bin": {
        "opencode": "./bin/opencode.cjs",
        "opencode2": "./bin/opencode2.cjs",
      },
      "dependencies": {
-        "@agentclientprotocol/sdk": "1.2.1",
+        "@agentclientprotocol/sdk": "1.6.0",
+        "@clack/core": "1.0.0-alpha.1",
        "@clack/prompts": "1.0.0-alpha.1",
        "@effect/platform-node": "catalog:",
        "@opencode-ai/pty": "0.1.13",
@@ -132,10 +158,11 @@
        "@opentui/solid": "catalog:",
        "@parcel/watcher": "2.5.1",
        "@silvia-odwyer/photon-node": "0.3.4",
+        "diff": "catalog:",
        "effect": "catalog:",
        "immer": "11.1.4",
        "jsonc-parser": "3.3.1",
-        "open": "10.1.2",
+        "picocolors": "1.1.1",
        "solid-js": "catalog:",
        "tree-sitter-bash": "0.25.0",
        "tree-sitter-powershell": "0.25.10",
@@ -177,7 +204,7 @@
    },
    "packages/client": {
      "name": "@opencode/client",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/protocol": "workspace:*",
        "@opencode/schema": "workspace:*",
@@ -203,7 +230,7 @@
    },
    "packages/codemode": {
      "name": "@opencode/codemode",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "acorn": "8.15.0",
        "effect": "catalog:",
@@ -216,7 +243,7 @@
    },
    "packages/console/app": {
      "name": "@opencode/console-app",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@cloudflare/vite-plugin": "1.15.2",
        "@ibm/plex": "6.4.1",
@@ -252,7 +279,7 @@
    },
    "packages/console/core": {
      "name": "@opencode/console-core",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@aws-sdk/client-sts": "3.782.0",
        "@jsx-email/render": "1.1.1",
@@ -279,7 +306,7 @@
    },
    "packages/console/function": {
      "name": "@opencode/console-function",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@openauthjs/openauth": "0.0.0-20250322224806",
        "@opencode/console-core": "workspace:*",
@@ -296,7 +323,7 @@
    },
    "packages/console/mail": {
      "name": "@opencode/console-mail",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@jsx-email/all": "2.2.3",
        "@jsx-email/cli": "1.4.3",
@@ -320,7 +347,7 @@
    },
    "packages/console/support": {
      "name": "@opencode/console-support",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@cloudflare/vite-plugin": "1.15.2",
        "@opencode/console-core": "workspace:*",
@@ -340,7 +367,7 @@
    },
    "packages/core": {
      "name": "@opencode/core",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@ai-sdk/cohere": "3.0.27",
        "@ai-sdk/gateway": "3.0.104",
@@ -376,6 +403,7 @@
        "https-proxy-agent": "7.0.6",
        "ignore": "7.0.5",
        "immer": "11.1.4",
+        "jose": "6.0.11",
        "jsonc-parser": "3.3.1",
        "mime-types": "3.0.2",
        "tree-sitter-bash": "0.25.0",
@@ -408,7 +436,7 @@
    },
    "packages/desktop": {
      "name": "@opencode/desktop",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@zip.js/zip.js": "2.7.62",
        "electron-context-menu": "5.0.0",
@@ -422,26 +450,22 @@
        "@lydell/node-pty": "catalog:",
        "@opencode/app": "workspace:*",
        "@opencode/client": "workspace:*",
-        "@opencode/plugin-browser": "workspace:*",
+        "@opencode/gui-extensions": "workspace:*",
        "@opencode/schema": "workspace:*",
        "@opencode/ui": "workspace:*",
        "@sentry/solid": "catalog:",
        "@sentry/vite-plugin": "catalog:",
-        "@solid-primitives/storage": "catalog:",
-        "@solidjs/meta": "catalog:",
        "@solidjs/router": "catalog:",
        "@types/bun": "catalog:",
        "@types/node": "catalog:",
        "@typescript/native-preview": "catalog:",
        "app-builder-lib": "26.15.7",
-        "devtools-protocol": "0.0.1687809",
        "drizzle-kit": "catalog:",
        "drizzle-orm": "catalog:",
        "effect": "catalog:",
-        "electron": "44.4.3",
+        "electron": "44.4.5",
        "electron-builder": "26.15.7",
        "electron-vite": "6.0.0-beta.1",
-        "puppeteer-core": "25.9.0",
        "solid-js": "catalog:",
        "typescript": "~5.6.2",
        "vite": "8.2.2",
@@ -457,7 +481,7 @@
    },
    "packages/enterprise": {
      "name": "@opencode/enterprise",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@hono/standard-validator": "catalog:",
        "@opencode-ai/sdk": "1.18.21",
@@ -494,7 +518,7 @@
    },
    "packages/function": {
      "name": "@opencode/function",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@octokit/auth-app": "8.0.1",
        "@octokit/rest": "catalog:",
@@ -508,9 +532,54 @@
        "typescript": "catalog:",
      },
    },
+    "packages/gui-extensions": {
+      "name": "@opencode/gui-extensions",
+      "version": "2.0.22",
+      "dependencies": {
+        "@dnd-kit/abstract": "0.5.0",
+        "@dnd-kit/dom": "0.5.0",
+        "@dnd-kit/solid": "0.5.0",
+        "@effect/platform-node": "catalog:",
+        "@kobalte/core": "catalog:",
+        "@opencode/client": "workspace:*",
+        "@opencode/plugin-browser": "workspace:*",
+        "@opencode/schema": "workspace:*",
+        "@opencode/session-ui": "workspace:*",
+        "@opencode/ui": "workspace:*",
+        "@opencode/util": "workspace:*",
+        "@pierre/trees": "1.0.0-beta.4",
+        "@solid-primitives/event-listener": "catalog:",
+        "@solid-primitives/keyed": "1.5.3",
+        "@solid-primitives/media": "catalog:",
+        "@solid-primitives/resize-observer": "catalog:",
+        "@solid-primitives/scheduled": "1.5.3",
+        "@tanstack/solid-query": "5.91.4",
+        "@tanstack/solid-virtual": "catalog:",
+        "effect": "catalog:",
+        "fuzzysort": "catalog:",
+        "ghostty-web": "github:anomalyco/ghostty-web#83c0a07b8628b748aed073b232cb4b52a6ca11c1",
+        "remeda": "catalog:",
+        "solid-js": "catalog:",
+        "solid-presence": "0.2.0",
+        "uqr": "0.1.3",
+      },
+      "devDependencies": {
+        "@happy-dom/global-registrator": "20.0.11",
+        "@lydell/node-pty": "catalog:",
+        "@tsconfig/node22": "catalog:",
+        "@types/bun": "catalog:",
+        "@types/node": "catalog:",
+        "@typescript/native-preview": "catalog:",
+        "devtools-protocol": "0.0.1687809",
+        "electron": "44.4.5",
+        "electron-updater": "6.8.9",
+        "lighthouse": "13.4.1",
+        "vite": "catalog:",
+      },
+    },
    "packages/http-recorder": {
      "name": "@opencode/http-recorder",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@effect/platform-node-shared": "4.0.0-rc.112",
      },
@@ -529,7 +598,7 @@
    },
    "packages/httpapi-codegen": {
      "name": "@opencode/httpapi-codegen",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "effect": "catalog:",
        "prettier": "3.6.2",
@@ -542,7 +611,7 @@
    },
    "packages/latex": {
      "name": "@opencode/latex",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/plugin": "workspace:*",
        "@opentui/core": "catalog:",
@@ -556,7 +625,7 @@
    },
    "packages/merman": {
      "name": "@opencode/merman",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/plugin": "workspace:*",
        "@opentui/core": "catalog:",
@@ -571,7 +640,7 @@
    },
    "packages/plugin": {
      "name": "@opencode/plugin",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@ai-sdk/provider": "3.0.8",
        "@opencode/ai": "workspace:*",
@@ -597,8 +666,8 @@
      },
      "peerDependencies": {
        "@opencode/theme": "workspace:*",
-        "@opentui/core": ">=0.5.12",
-        "@opentui/solid": ">=0.5.12",
+        "@opentui/core": ">=0.5.14",
+        "@opentui/solid": ">=0.5.14",
        "solid-js": ">=1.9.0",
      },
      "optionalPeers": [
@@ -610,7 +679,7 @@
    },
    "packages/plugin-browser": {
      "name": "@opencode/plugin-browser",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/plugin": "workspace:*",
        "@opencode/schema": "workspace:*",
@@ -640,7 +709,7 @@
    },
    "packages/protocol": {
      "name": "@opencode/protocol",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/schema": "workspace:*",
        "effect": "catalog:",
@@ -655,7 +724,7 @@
    },
    "packages/schema": {
      "name": "@opencode/schema",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@standard-schema/spec": "catalog:",
        "effect": "catalog:",
@@ -679,7 +748,7 @@
    },
    "packages/sdk": {
      "name": "@opencode/sdk",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/client": "workspace:*",
        "@opencode/core": "workspace:*",
@@ -700,7 +769,7 @@
    },
    "packages/server": {
      "name": "@opencode/server",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@effect/platform-node": "catalog:",
        "@effect/platform-node-shared": "catalog:",
@@ -722,7 +791,7 @@
    },
    "packages/session-ui": {
      "name": "@opencode/session-ui",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@kobalte/core": "catalog:",
        "@opencode/client": "workspace:*",
@@ -757,7 +826,7 @@
    },
    "packages/simulation": {
      "name": "@opencode/simulation",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/ai": "workspace:*",
        "@opencode/core": "workspace:*",
@@ -777,7 +846,7 @@
    },
    "packages/stats/app": {
      "name": "@opencode/stats-app",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@ibm/plex": "6.4.1",
        "@kobalte/core": "catalog:",
@@ -811,7 +880,7 @@
    },
    "packages/stats/core": {
      "name": "@opencode/stats-core",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@aws-sdk/client-athena": "3.933.0",
        "@planetscale/database": "1.19.0",
@@ -830,7 +899,7 @@
    },
    "packages/stats/server": {
      "name": "@opencode/stats-server",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@aws-sdk/client-firehose": "3.933.0",
        "@effect/platform-node": "catalog:",
@@ -876,7 +945,7 @@
    },
    "packages/theme": {
      "name": "@opencode/theme",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opentui/core": "catalog:",
        "effect": "catalog:",
@@ -890,7 +959,7 @@
    },
    "packages/tui": {
      "name": "@opencode/tui",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@opencode/client": "workspace:*",
        "@opencode/core": "workspace:*",
@@ -909,7 +978,6 @@
        "effect": "catalog:",
        "fuzzysort": "catalog:",
        "get-east-asian-width": "catalog:",
-        "open": "10.1.2",
        "opentui-spinner": "catalog:",
        "remeda": "catalog:",
        "solid-js": "catalog:",
@@ -925,7 +993,7 @@
    },
    "packages/ui": {
      "name": "@opencode/ui",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@kobalte/core": "catalog:",
        "@pierre/diffs": "catalog:",
@@ -960,7 +1028,7 @@
    },
    "packages/util": {
      "name": "@opencode/util",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@effect/opentelemetry": "catalog:",
        "@effect/platform-node": "catalog:",
@@ -982,6 +1050,7 @@
        "mime-types": "3.0.2",
        "minimatch": "10.2.5",
        "npm-package-arg": "13.0.2",
+        "open": "11.0.4",
        "pacote": "21.5.1",
      },
      "devDependencies": {
@@ -997,7 +1066,7 @@
    },
    "packages/web": {
      "name": "@opencode/web",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "@astrojs/cloudflare": "12.6.3",
        "@astrojs/markdown-remark": "6.3.1",
@@ -1038,7 +1107,7 @@
    },
    "services/update": {
      "name": "@opencode/update",
-      "version": "2.0.16",
+      "version": "2.0.22",
      "dependencies": {
        "jose": "6.0.11",
        "semver": "catalog:",
@@ -1098,6 +1167,7 @@
    "@types/node": "catalog:",
    "bun-types": "1.4.2",
    "effect": "catalog:",
+    "open": "11.0.4",
    "solid-js": "catalog:",
  },
  "catalog": {
@@ -1114,10 +1184,10 @@
    "@npmcli/arborist": "9.4.0",
    "@octokit/rest": "22.0.0",
    "@openauthjs/openauth": "0.0.0-20250322224806",
-    "@opentui/core": "0.5.12",
-    "@opentui/keymap": "0.5.12",
-    "@opentui/solid": "0.5.12",
-    "@pierre/diffs": "1.2.10",
+    "@opentui/core": "0.5.14",
+    "@opentui/keymap": "0.5.14",
+    "@opentui/solid": "0.5.14",
+    "@pierre/diffs": "1.5.1",
    "@playwright/test": "1.59.1",
    "@sentry/solid": "10.71.0",
    "@sentry/vite-plugin": "5.4.0",
@@ -1186,7 +1256,7 @@

    "@adobe/css-tools": ["@adobe/css-tools@4.5.0", "", {}, "sha512-6OzddxPio9UiWTCemp4N8cYLV2ZN1ncRnV1cVGtve7dhPOtRkleRyx32GQCYSwDYgaHU3USMm84tNsvKzRCa1Q=="],

-    "@agentclientprotocol/sdk": ["@agentclientprotocol/sdk@1.2.1", "", { "peerDependencies": { "zod": "^3.25.0 || ^4.0.0" } }, "sha512-jwYUdOQR7tc+Zfch53VL4JJyUNK/46q03uUTYb+PjECsmnNl94XFXOfYLJ8RBpMNidXd1rpOAVgb0vqD98xImA=="],
+    "@agentclientprotocol/sdk": ["@agentclientprotocol/sdk@1.6.0", "", { "peerDependencies": { "zod": "^3.25.0 || ^4.0.0" } }, "sha512-XxXrmX7aZkDgOB0Rg9cu+ZFyiUUc5lF2n9seO3Gc4OR+MTdfZOwIqF6m3LvsmmM8K3qgPmXkDH8/IFM2u9vdcQ=="],

    "@ai-sdk/cohere": ["@ai-sdk/cohere@3.0.27", "", { "dependencies": { "@ai-sdk/provider": "3.0.8", "@ai-sdk/provider-utils": "4.0.21" }, "peerDependencies": { "zod": "^3.25.76 || ^4.1.8" } }, "sha512-OqcCq2PiFY1dbK/0Ck45KuvE8jfdxRuuAE9Y5w46dAk6U+9vPOeg1CDcmR+ncqmrYrhRl3nmyDttyDahyjCzAw=="],

@@ -2144,6 +2214,8 @@

    "@opencode/app": ["@opencode/app@workspace:packages/app"],

+    "@opencode/browser-extension": ["@opencode/browser-extension@workspace:packages/browser-extension"],
+
    "@opencode/cli": ["@opencode/cli@workspace:packages/cli"],

    "@opencode/client": ["@opencode/client@workspace:packages/client"],
@@ -2172,6 +2244,8 @@

    "@opencode/function": ["@opencode/function@workspace:packages/function"],

+    "@opencode/gui-extensions": ["@opencode/gui-extensions@workspace:packages/gui-extensions"],
+
    "@opencode/http-recorder": ["@opencode/http-recorder@workspace:packages/http-recorder"],

    "@opencode/httpapi-codegen": ["@opencode/httpapi-codegen@workspace:packages/httpapi-codegen"],
@@ -2252,27 +2326,27 @@

    "@opentelemetry/semantic-conventions": ["@opentelemetry/semantic-conventions@1.43.0", "", {}, "sha512-eSYWTm620tTk45EKSedaUL8MFYI8hW164hIXsgIHyxu3VobUB3fFCu5t0hQby6OoWRPsG1KkKUG2M5UadiLiVg=="],

-    "@opentui/core": ["@opentui/core@0.5.12", "", { "dependencies": { "bun-ffi-structs": "0.3.1", "diff": "9.0.0", "marked": "17.0.1", "string-width": "7.2.0", "strip-ansi": "7.1.2" }, "optionalDependencies": { "@opentui/core-darwin-arm64": "0.5.12", "@opentui/core-darwin-x64": "0.5.12", "@opentui/core-linux-arm64": "0.5.12", "@opentui/core-linux-arm64-musl": "0.5.12", "@opentui/core-linux-x64": "0.5.12", "@opentui/core-linux-x64-musl": "0.5.12", "@opentui/core-win32-arm64": "0.5.12", "@opentui/core-win32-x64": "0.5.12" }, "peerDependencies": { "web-tree-sitter": "0.25.10" } }, "sha512-ZXBE5gmvdovmV8zJQrOQf6E44v1tJRDEgrM2MYhEglzgXZ+smIUp95O8zeRYGsuIzQIiMPMgQqKtTJuzvAb7BQ=="],
+    "@opentui/core": ["@opentui/core@0.5.14", "", { "dependencies": { "bun-ffi-structs": "0.3.1", "diff": "9.0.0", "marked": "17.0.1", "string-width": "7.2.0", "strip-ansi": "7.1.2" }, "optionalDependencies": { "@opentui/core-darwin-arm64": "0.5.14", "@opentui/core-darwin-x64": "0.5.14", "@opentui/core-linux-arm64": "0.5.14", "@opentui/core-linux-arm64-musl": "0.5.14", "@opentui/core-linux-x64": "0.5.14", "@opentui/core-linux-x64-musl": "0.5.14", "@opentui/core-win32-arm64": "0.5.14", "@opentui/core-win32-x64": "0.5.14" }, "peerDependencies": { "web-tree-sitter": "0.25.10" } }, "sha512-tfQ+PWQyeBnYloB3diEcPqbILv14xemH5jjAEICfPuyNDtGBqrjhUtThrhbvFRuPMcj6IEeXrAk6VE8e91h0kg=="],

-    "@opentui/core-darwin-arm64": ["@opentui/core-darwin-arm64@0.5.12", "", { "os": "darwin", "cpu": "arm64" }, "sha512-YdVnP0tAyerBNl0mIcmQEOotPeZzW1VnSXKBl5cyZ5e6nDd2Y+ui/8eRPpn1oqcamf1NCnzS4ohMgejOvna8Zg=="],
+    "@opentui/core-darwin-arm64": ["@opentui/core-darwin-arm64@0.5.14", "", { "os": "darwin", "cpu": "arm64" }, "sha512-wWmw41wRMBoI0lN9mgyzGRwYujwwfNkBP6jYM99k4Bx2XkeSZ3sCeu5bwX9vEJPYqjiBShbDhc0NnVSP7NwzVg=="],

-    "@opentui/core-darwin-x64": ["@opentui/core-darwin-x64@0.5.12", "", { "os": "darwin", "cpu": "x64" }, "sha512-uRrQJdHmLUSj3PV23QPi3WSimYTTxcXnVouxF6U4xMXlOv4N3SxnHfVwMRQkPqbGOfvVWHeLE6FdK4C+ubU0sQ=="],
+    "@opentui/core-darwin-x64": ["@opentui/core-darwin-x64@0.5.14", "", { "os": "darwin", "cpu": "x64" }, "sha512-7smHKDH8IhUaBsgYuAClMl2mHWu+yjpMrtdw3wEvBVtCzKMrHi/TJ5PtydO+IfUzR3BXpVubdbR1irD8BTcR/w=="],

-    "@opentui/core-linux-arm64": ["@opentui/core-linux-arm64@0.5.12", "", { "os": "linux", "cpu": "arm64" }, "sha512-XeKhuIaEtgipvuPHbl4qPOBj+Ut+2zObmsxMVM1jDcjz/FatG9PGeGQPx1G1SnvH2AgpT4K+eCu7DUF0+yIqoQ=="],
+    "@opentui/core-linux-arm64": ["@opentui/core-linux-arm64@0.5.14", "", { "os": "linux", "cpu": "arm64" }, "sha512-xH1hP+NaLySEJeZkl21NlkZBMddMfQ1jU8NeX1AEBc2GNBOvDXU4Ud/xw87SrAvU1xG9K7/9C4oy4AmIMEpVGg=="],

-    "@opentui/core-linux-arm64-musl": ["@opentui/core-linux-arm64-musl@0.5.12", "", { "os": "linux", "cpu": "arm64" }, "sha512-VZ2sNMw1d/r1SLPjUbOP9LKscKz1CQjID8adTL6gG8Lrrq+mYcIUxutyB+P/eG0J/7oRZLPR6OMt7dUOap6RTg=="],
+    "@opentui/core-linux-arm64-musl": ["@opentui/core-linux-arm64-musl@0.5.14", "", { "os": "linux", "cpu": "arm64" }, "sha512-mnBBAuTb92NiRLAjOD755tS8/tNQemDztbg9tMvoCT90G52FtVrRb31Ge6OrYqfm0c9DkZGhEBOhunsId/4zSA=="],

-    "@opentui/core-linux-x64": ["@opentui/core-linux-x64@0.5.12", "", { "os": "linux", "cpu": "x64" }, "sha512-eZiCjEzwbb6qClPPfk32Nha9xmr9obt69Xj0+9SKsXxWLBKkjQEGOMRoh/R9ObaQF4aq8If1xV3VEY0sD9W9vg=="],
+    "@opentui/core-linux-x64": ["@opentui/core-linux-x64@0.5.14", "", { "os": "linux", "cpu": "x64" }, "sha512-Hkk4kaDGMcn9bmJFJPW3/QOGGbPuWe3sCFV/CkiVb4+bCvcTm7EXQr4QTAA63BYy2dKE5bUFUi1zZlyeMkWpnA=="],

-    "@opentui/core-linux-x64-musl": ["@opentui/core-linux-x64-musl@0.5.12", "", { "os": "linux", "cpu": "x64" }, "sha512-WWW0hVBoSYZ3D6AgZ4u2Y5/u/IyIq2pDb+4yI3WgJ70Wyt6ofHy+6kRGRgbXFn1p+rPInAHjCXD2v6C7iEKSrA=="],
+    "@opentui/core-linux-x64-musl": ["@opentui/core-linux-x64-musl@0.5.14", "", { "os": "linux", "cpu": "x64" }, "sha512-ngJ+U2grOGEteeQvZdAJNqn09M+At5mfWInauK5aS427bea1yLo+e6hor/CRmbn9SxEEF+SwoyekaOrLPWyU7w=="],

-    "@opentui/core-win32-arm64": ["@opentui/core-win32-arm64@0.5.12", "", { "os": "win32", "cpu": "arm64" }, "sha512-aLbm6870Ybls6CYL4zMOCImTBPLZHZMUXJFGqMI44lIWxitkAtT6zg5lYA4oRqFRzzryDclxr29+hDgT3p3Blw=="],
+    "@opentui/core-win32-arm64": ["@opentui/core-win32-arm64@0.5.14", "", { "os": "win32", "cpu": "arm64" }, "sha512-T9kNqKXg2jysmTsyyZ1A8LBQotFBM+iPjzyRslxyexqrs8a1UcmxbApEo+COtxQuqukMTbvwyqEdAT8vcmxEkQ=="],

-    "@opentui/core-win32-x64": ["@opentui/core-win32-x64@0.5.12", "", { "os": "win32", "cpu": "x64" }, "sha512-KTwtwpfd2zF9opVh3SyRJYDd1o3Xv4XL8OZb8Zi+CqWUel6Y2IDCiVivCv8fGJt3J7wOIXXtuZI9ZUkLyKJCiQ=="],
+    "@opentui/core-win32-x64": ["@opentui/core-win32-x64@0.5.14", "", { "os": "win32", "cpu": "x64" }, "sha512-mqKSkab8VdMLSmMdocna7+BTyMoIutkVXOV9lfmPrBO2g7Np5c6c4SJ4QIVZqPvast11XyT0/7FfXOYLKAS72w=="],

-    "@opentui/keymap": ["@opentui/keymap@0.5.12", "", { "dependencies": { "@opentui/core": "0.5.12" }, "peerDependencies": { "@opentui/react": "0.5.12", "@opentui/solid": "0.5.12", "react": ">=19.2.0", "solid-js": "1.9.12" }, "optionalPeers": ["@opentui/react", "@opentui/solid", "react", "solid-js"] }, "sha512-yWPvJjRhJTRoRSUucQq9Ua8ZW7n/2YQ/j6JxWq5Qekm4WuFiTplEkebR/Aj2/xA8tX68NOE5qv1LrY0Jk3NLNQ=="],
+    "@opentui/keymap": ["@opentui/keymap@0.5.14", "", { "dependencies": { "@opentui/core": "0.5.14" }, "peerDependencies": { "@opentui/react": "0.5.14", "@opentui/solid": "0.5.14", "react": ">=19.2.0", "solid-js": "1.9.12" }, "optionalPeers": ["@opentui/react", "@opentui/solid", "react", "solid-js"] }, "sha512-YGTAvRrpQTbRNV7GH0UxRuCSPxwnSooVdB+qPHHP+3ywc93+lhekljgP8dRivl32f3CiMqqm96Uj2PgwKxkC0Q=="],

-    "@opentui/solid": ["@opentui/solid@0.5.12", "", { "dependencies": { "@babel/core": "7.28.0", "@babel/preset-typescript": "7.27.1", "@opentui/core": "0.5.12", "babel-plugin-module-resolver": "5.0.2", "babel-preset-solid": "1.9.12", "entities": "7.0.1", "s-js": "^0.4.9" }, "peerDependencies": { "solid-js": "1.9.12" } }, "sha512-hAiVlVMtT7AkHGblKwcW1YAuXtxkSy1XSf/RRc4j3IlG3mTNX0bhJdnGOo3Xw14EqeZMp41Mcp5WzHAzMm/DzA=="],
+    "@opentui/solid": ["@opentui/solid@0.5.14", "", { "dependencies": { "@babel/core": "7.28.0", "@babel/preset-typescript": "7.27.1", "@opentui/core": "0.5.14", "babel-plugin-module-resolver": "5.0.2", "babel-preset-solid": "1.9.12", "entities": "7.0.1", "s-js": "^0.4.9" }, "peerDependencies": { "solid-js": "1.9.12" } }, "sha512-bBRl34mZ0wFGhjHX6y1VNiZmJ3DSm2DVPm0PNSVDpvq9qxRHuywEPTCn/Lgte0C9450tgf11cDqcE3pR9cZwVA=="],

    "@oslojs/asn1": ["@oslojs/asn1@1.0.0", "", { "dependencies": { "@oslojs/binary": "1.0.0" } }, "sha512-zw/wn0sj0j0QKbIXfIlnEcTviaCzYOY3V5rAyjR6YtOByFtJiT574+8p9Wlach0lZH9fddD4yb9laEAIl4vXQA=="],

@@ -2474,6 +2548,8 @@

    "@oxlint/binding-win32-x64-msvc": ["@oxlint/binding-win32-x64-msvc@1.60.0", "", { "os": "win32", "cpu": "x64" }, "sha512-JOro4ZcfBLamJCyfURQmOQByoorgOdx3ZjAkSqnb/CyG/i+lN3KoV5LAgk5ZAW6DPq7/Cx7n23f8DuTWXTWgyQ=="],

+    "@oxlint/plugins": ["@oxlint/plugins@1.60.0", "", {}, "sha512-wxEoVVAS5FQAXiA8jPY+NWTOTKTzbSViMfW7yU5OQ7+f34W0KYfy2iuSLC9o1fiYn5xt3LQLxZBREFJtDlQLRw=="],
+
    "@pagefind/darwin-arm64": ["@pagefind/darwin-arm64@1.5.2", "", { "os": "darwin", "cpu": "arm64" }, "sha512-MXpI+7HsAdPkvJ0gk9xj9g541BCqBZOBbdwj9g6lB5LCj6kSV6nqDSjzcAJwvOsfu0fjwvC8hQU+ecfhp+MpiQ=="],

    "@pagefind/darwin-x64": ["@pagefind/darwin-x64@1.5.2", "", { "os": "darwin", "cpu": "x64" }, "sha512-IojxFWMEJe0RQ7PQ3KXQsPIImNsbpPYpoZ+QUDrL8fAl/O27IX+LVLs74/UzEZy5uA2LD8Nz1AiwKr72vrkZQw=="],
@@ -2528,11 +2604,11 @@

    "@peculiar/webcrypto": ["@peculiar/webcrypto@1.7.1", "", { "dependencies": { "@peculiar/asn1-schema": "^2.7.0", "@peculiar/json-schema": "^1.1.12", "@peculiar/utils": "^2.0.2", "tslib": "^2.8.1", "webcrypto-core": "^1.9.2" } }, "sha512-ODOov0sGMJMf3jPonOkgGqPknTsu+DdQ7kD++gz8aI+aFMOMHFbWAA2taqXXVTdP+OTOQR/znGvSpmkeI0WTYQ=="],

-    "@pierre/diffs": ["@pierre/diffs@1.2.10", "", { "dependencies": { "@pierre/theme": "1.0.3", "@pierre/theming": "0.0.1", "@shikijs/transformers": "^3.0.0 || ^4.0.0", "diff": "8.0.3", "hast-util-to-html": "9.0.5", "lru_map": "0.4.1", "shiki": "^3.0.0 || ^4.0.0" }, "peerDependencies": { "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0" } }, "sha512-rPeAmDWarxFVTQpaf4y6wTxjZxU44xKJKoJti2zU21P06DVd9nRHZX+xSIObLB307Qjpaesyb1x/j0z94t7vLw=="],
+    "@pierre/diffs": ["@pierre/diffs@1.5.1", "", { "dependencies": { "@pierre/theme": "2.0.0", "@pierre/theming": "1.0.1", "@shikijs/transformers": "^3.0.0 || ^4.0.0", "diff": "9.0.0", "hast-util-to-html": "9.0.5", "lru_map": "0.4.1", "shiki": "^3.0.0 || ^4.0.0" }, "peerDependencies": { "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0" }, "optionalPeers": ["react", "react-dom"] }, "sha512-+EXNfz4ZXI6FHr7P6ToWEORwyYiPVfAmzhA70xF+0y/oNvj7gbNTLJJ6jNzkBjuMbRlu564TAmgYZLc5OyHURA=="],

-    "@pierre/theme": ["@pierre/theme@1.0.3", "", {}, "sha512-sWHv11TMoqKxKDgTIk5VbhQjdPhs8DCcBxbjh3mRlS3YOM/OcrWoGX6MM8eBGn9cUu3M46Py0JnxsG2nJaFTuA=="],
+    "@pierre/theme": ["@pierre/theme@2.0.0", "", {}, "sha512-yNDd9GYLQl1mEUJR8AneJ5e4ohLIHQd/wZLWr4fagt78vS2RwwZNW530vVgHqXFAyFVcFlRmGUD5ramXH46OXw=="],

-    "@pierre/theming": ["@pierre/theming@0.0.1", "", { "peerDependencies": { "@pierre/theme": "^1.0.0", "@shikijs/themes": "^3.0.0 || ^4.0.0", "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0", "shiki": "^3.0.0 || ^4.0.0" }, "optionalPeers": ["@pierre/theme", "@shikijs/themes", "react", "react-dom", "shiki"] }, "sha512-1thlEtJbqdyLzc1ZS2KQa1q7FzDGHT4dTEdKHoyQjOMeWWOmbVG5/ndEfOKfAb5Fzkz8cNJrOjFLiZoDH/A03A=="],
+    "@pierre/theming": ["@pierre/theming@1.0.1", "", { "peerDependencies": { "@pierre/theme": "^1.1.0 || ^2.0.0", "@shikijs/themes": "^3.0.0 || ^4.0.0", "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0", "shiki": "^3.0.0 || ^4.0.0" }, "optionalPeers": ["@pierre/theme", "@shikijs/themes", "react", "react-dom", "shiki"] }, "sha512-WCI5Qd7iprDpISL9fBYOLe8RV53+b7mFNA3bPzl60/2CKCSrsKN8zEcep6Y3BAzvARlmca50zGjDodqPGiTUKA=="],

    "@pierre/trees": ["@pierre/trees@1.0.0-beta.4", "", { "dependencies": { "preact": "11.0.0-beta.0", "preact-render-to-string": "6.6.5" }, "peerDependencies": { "react": "^18.3.1 || ^19.0.0", "react-dom": "^18.3.1 || ^19.0.0" } }, "sha512-OfT1yk9ne8Te5+GB5zUY8yqE6B8BqjBHQJleH4lu8ltwNpoocZl4vXt1AzlEExpxI/pp+AFX5QG+lR3JjtTEag=="],

@@ -3062,6 +3138,8 @@

    "@types/chai": ["@types/chai@5.2.3", "", { "dependencies": { "@types/deep-eql": "*", "assertion-error": "^2.0.1" } }, "sha512-Mw558oeA9fFbv65/y4mHtXDs9bPnFMZAL/jxdPFUpOHHIXX91mcgEHbS5Lahr+pwZFR8A7GQleRWeI6cGFC2UA=="],

+    "@types/chrome": ["@types/chrome@0.3.4", "", { "dependencies": { "@types/filesystem": "*", "@types/har-format": "*" } }, "sha512-dcySM5R3WAUVYFjylu6tvI4D03hPKjgAqk+tKZ/tm/vxL4mLgbnSsOEQ3J7V99qP/U3Oo9IORstJLmX2hPaDlA=="],
+
    "@types/cross-spawn": ["@types/cross-spawn@6.0.6", "", { "dependencies": { "@types/node": "*" } }, "sha512-fXRhhUkG4H3TQk5dBhQ7m/JDdSNHKwR2BBia62lhwEIq9xGiQKLxd6LymNhn47SjXhsUEPmxi+PKw2OkW4LLjA=="],

    "@types/d3": ["@types/d3@7.4.3", "", { "dependencies": { "@types/d3-array": "*", "@types/d3-axis": "*", "@types/d3-brush": "*", "@types/d3-chord": "*", "@types/d3-color": "*", "@types/d3-contour": "*", "@types/d3-delaunay": "*", "@types/d3-dispatch": "*", "@types/d3-drag": "*", "@types/d3-dsv": "*", "@types/d3-ease": "*", "@types/d3-fetch": "*", "@types/d3-force": "*", "@types/d3-format": "*", "@types/d3-geo": "*", "@types/d3-hierarchy": "*", "@types/d3-interpolate": "*", "@types/d3-path": "*", "@types/d3-polygon": "*", "@types/d3-quadtree": "*", "@types/d3-random": "*", "@types/d3-scale": "*", "@types/d3-scale-chromatic": "*", "@types/d3-selection": "*", "@types/d3-shape": "*", "@types/d3-time": "*", "@types/d3-time-format": "*", "@types/d3-timer": "*", "@types/d3-transition": "*", "@types/d3-zoom": "*" } }, "sha512-lZXZ9ckh5R8uiFVt8ogUNf+pIrK4EsWrx2Np75WvF/eTpJ0FMHNhjXk8CKEx/+gpHbNQyJWehbFaTvqmHWB3ww=="],
@@ -3134,12 +3212,18 @@

    "@types/estree-jsx": ["@types/estree-jsx@1.0.5", "", { "dependencies": { "@types/estree": "*" } }, "sha512-52CcUVNFyfb1A2ALocQw/Dd1BQFNmSdkuC3BkZ6iqhdMfQz7JWOFRuJFloOzjk+6WijU56m9oKXFAXc7o3Towg=="],

+    "@types/filesystem": ["@types/filesystem@0.0.36", "", { "dependencies": { "@types/filewriter": "*" } }, "sha512-vPDXOZuannb9FZdxgHnqSwAG/jvdGM8Wq+6N4D/d80z+D4HWH+bItqsZaVRQykAn6WEVeEkLm2oQigyHtgb0RA=="],
+
+    "@types/filewriter": ["@types/filewriter@0.0.33", "", {}, "sha512-xFU8ZXTw4gd358lb2jw25nxY9QAgqn2+bKKjKOYfNCzN4DKCFetK7sPtrlpg66Ywe3vWY9FNxprZawAh9wfJ3g=="],
+
    "@types/fontkit": ["@types/fontkit@2.0.9", "", { "dependencies": { "@types/node": "*" } }, "sha512-qNYerFky3muCmZPq+R+B3cUDRA5OONw/oh6aGGFxx2LOBz6yu8eamKusrhkHnC6rc2fm76+G9z9QoWSB2SaQaw=="],

    "@types/fs-extra": ["@types/fs-extra@9.0.13", "", { "dependencies": { "@types/node": "*" } }, "sha512-nEnwB++1u5lVDM2UI4c1+5R+FYaKfaAzS4OococimjVm3nQw3TuzH5UNsocrcTBbhnerblyHj4A49qXbIiZdpA=="],

    "@types/geojson": ["@types/geojson@7946.0.16", "", {}, "sha512-6C8nqWur3j98U6+lXDfTUWIfgvZU+EumvpHKcYjujKH7woYyLj2sUmff0tRhrqM7BohUw7Pz3ZB1jj2gW9Fvmg=="],

+    "@types/har-format": ["@types/har-format@1.2.16", "", {}, "sha512-fluxdy7ryD3MV6h8pTfTYpy/xQzCFC7m89nOH9y94cNqJ1mDIDPut7MnRHI3F6qRmh/cT2fUjG1MLdCNb4hE9A=="],
+
    "@types/hast": ["@types/hast@3.0.5", "", { "dependencies": { "@types/unist": "*" } }, "sha512-rp/ezSWaD1m44dPKICGhiskI13nVr7qTloFwDa/IYkhhf5nzwP+zIQcIJh3WIFSBOy/H1PzB40jPjMDksN4F+g=="],

    "@types/http-cache-semantics": ["@types/http-cache-semantics@4.2.0", "", {}, "sha512-L3LgimLHXtGkWikKnsPg0/VFx9OGZaC+eN1u4r+OB1XRqH3meBIAVC2zr1WdMH+RHmnRkqliQAOHNJ/E0j/e0Q=="],
@@ -3868,7 +3952,7 @@

    "ejs": ["ejs@3.1.10", "", { "dependencies": { "jake": "^10.8.5" }, "bin": { "ejs": "bin/cli.js" } }, "sha512-UeJmFfOrAQS8OJWPZ4qtgHyWExa088/MtK5UEyoJGFH67cDEXkZSviOiKRCZ4Xij0zxI3JECgYs3oKx+AizQBA=="],

-    "electron": ["electron@44.4.3", "", { "dependencies": { "@electron-internal/extract-zip": "^1.0.1", "@electron/get": "^5.0.0", "@types/node": "^24.9.0" }, "bin": { "electron": "cli.js", "install-electron": "install.js" } }, "sha512-LTpSFTB40qVCXIX5xMo+cgHI/Jjkbjw7VpB26PccEbroqOn72LBukeaDwPVo1fBYzSzs0c9iPuAucFCO7Tw81Q=="],
+    "electron": ["electron@44.4.5", "", { "dependencies": { "@electron-internal/extract-zip": "^1.0.1", "@electron/get": "^5.0.0", "@types/node": "^24.9.0" }, "bin": { "electron": "cli.js", "install-electron": "install.js" } }, "sha512-SjgoaeYsSWZfJzubgQU7juvuXMTvn6/e1gAHdGFA/yuMbpF+I+skYqIJ6DdXBXHeWbkbpNpmwZCTpiivtlWZSw=="],

    "electron-builder": ["electron-builder@26.15.7", "", { "dependencies": { "app-builder-lib": "26.15.7", "builder-util": "26.15.3", "builder-util-runtime": "9.7.0", "chalk": "^4.1.2", "ci-info": "^4.2.0", "dmg-builder": "26.15.7", "fs-extra": "^10.1.0", "lazy-val": "^1.0.5", "simple-update-notifier": "2.0.0", "yargs": "^17.6.2" }, "bin": { "electron-builder": "./cli.js", "install-app-deps": "./install-app-deps.js" } }, "sha512-DBpaNzxsPs1BvEblzFoNriSbzsBqDCy/gseIngeEhYzQG1IxfB7Hvc2tBBVmpWE2BTQGP9J1RrAvDT+Vc/uAxg=="],

@@ -4322,7 +4406,7 @@

    "is-decimal": ["is-decimal@2.0.1", "", {}, "sha512-AAB9hiomQs5DXWcRB1rqsxGUstbRroFOPPVAomNk/3XHR5JyEZChOyTWe2oayKnsSsr/kcGqF+z6yuH6HHpN0A=="],

-    "is-docker": ["is-docker@3.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-eljcgEDlEns/7AXFosB5K/2nCM4P7FQPkGc/DWLy5rmFEWvZayGrik1d9/QIY5nJ4f9YsVvBkA6kJpHn9rISdQ=="],
+    "is-docker": ["is-docker@4.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-LHE+wROyG/Y/0ZnbktRCoTix2c1RhgWaZraMZ8o1Q7zCh0VSrICJQO5oqIIISrcSBtrXv0o233w1IYwsWCjTzA=="],

    "is-document.all": ["is-document.all@1.0.0", "", { "dependencies": { "call-bound": "^1.0.4" } }, "sha512-+XSoyS05OdBbhFuELhgTCpFNHkpBOJqtsZfUFFpe5QTw+9Sjbh8zitxhQkYAo6wV7e1Vb8cAPvpCk9jGam/82g=="],

@@ -4340,6 +4424,8 @@

    "is-hexadecimal": ["is-hexadecimal@2.0.1", "", {}, "sha512-DgZQp241c8oO6cA1SbTEWiXeoxV42vlcJxgH+B3hi1AiqqKruZR3ZGF8In3fj4+/y/7rHvlOZLZtgJ/4ttYGZg=="],

+    "is-in-ssh": ["is-in-ssh@1.0.0", "", {}, "sha512-jYa6Q9rH90kR1vKB6NM7qqd1mge3Fx4Dhw5TVlK1MUBqhEOuCagrEHMevNuCcbECmXZ0ThXkRm+Ymr51HwEPAw=="],
+
    "is-inside-container": ["is-inside-container@1.0.0", "", { "dependencies": { "is-docker": "^3.0.0" }, "bin": { "is-inside-container": "cli.js" } }, "sha512-KIYLCCJghfHZxqjYBE7rEy0OBuTd5xCHS7tHVgvCLkx7StIoaxwNW3hCALgEUjFfeRk+MG/Qxmp/vtETEF3tRA=="],

    "is-map": ["is-map@2.0.3", "", {}, "sha512-1Qed0/Hr2m+YqxnM09CjA2d/i6YZNfF6R2oRAOj36eUdS6qIV/huPJNSEpKbupewFs+ZsJlxsjjPbc0/afW6Lw=="],
@@ -4386,7 +4472,7 @@

    "is-whitespace": ["is-whitespace@0.3.0", "", {}, "sha512-RydPhl4S6JwAyj0JJjshWJEFG6hNye3pZFBRZaTUfZFwGHxzppNaNOVgQuS/E/SlhrApuMXrpnK1EEIXfdo3Dg=="],

-    "is-wsl": ["is-wsl@3.1.1", "", { "dependencies": { "is-inside-container": "^1.0.0" } }, "sha512-e6rvdUCiQCAuumZslxRJWR/Doq4VpPR82kqclvcS0efgt430SlGIk05vdCN58+VrzgtIcfNODjozVielycD4Sw=="],
+    "is-wsl": ["is-wsl@2.2.0", "", { "dependencies": { "is-docker": "^2.0.0" } }, "sha512-fKzAra0rGJUUBwGBgNkHZuToZcn+TtXHpeCgmkMJMMYx1sQDYaCSyjJBSCa2nH1DGm7s3n1oBnohoVTBaN7Lww=="],

    "isarray": ["isarray@1.0.0", "", {}, "sha512-VLghIWNM6ELQzo7zwmcg0NmTVyWKYjvIeM83yjp0wRDTmUnrM678fQbcKBo6n2CJEF0szoG//ytg+TKla89ALQ=="],

@@ -4850,7 +4936,7 @@

    "oniguruma-to-es": ["oniguruma-to-es@4.3.6", "", { "dependencies": { "oniguruma-parser": "^0.12.2", "regex": "^6.1.0", "regex-recursion": "^6.0.2" } }, "sha512-csuQ9x3Yr0cEIs/Zgx/OEt9iBw9vqIunAPQkx19R/fiMq2oGVTgcMqO/V3Ybqefr1TBvosI6jU539ksaBULJyA=="],

-    "open": ["open@10.1.2", "", { "dependencies": { "default-browser": "^5.2.1", "define-lazy-prop": "^3.0.0", "is-inside-container": "^1.0.0", "is-wsl": "^3.1.0" } }, "sha512-cxN6aIDPz6rm8hbebcP7vrQNhvRcveZoJU72Y7vskh4oIm+BZwBECnx5nTmrlres1Qapvx27Qo1Auukpf8PKXw=="],
+    "open": ["open@11.0.4", "", { "dependencies": { "default-browser": "^5.5.1", "define-lazy-prop": "^3.0.0", "is-in-ssh": "^1.0.0", "is-inside-container": "^1.0.0", "powershell-utils": "^0.2.1", "wsl-utils": "^1.0.0" } }, "sha512-++Zlftm0kVLPmzC06t6epuWmcRMDbI4z5P3NNX979WA/k23+NtSOynEGzsVfZwguKw2mi5umVgnBlJQMwRz4Pg=="],

    "openai": ["openai@6.49.0", "", { "peerDependencies": { "@aws-sdk/credential-provider-node": ">=3.972.0 <4", "@smithy/hash-node": ">=4.3.0 <5", "@smithy/signature-v4": ">=5.4.0 <6", "ws": "^8.18.0", "zod": "^3.25 || ^4.0" }, "optionalPeers": ["@aws-sdk/credential-provider-node", "@smithy/hash-node", "@smithy/signature-v4", "ws", "zod"] }, "sha512-aYCc0C6L864eR6WSYIwQGyXriw/nIyZx0ObvhzOEVuk0zoBDpynjSbrionWI7q65B5H8jJX0DXR9snEzM6bfPg=="],

@@ -4994,6 +5080,8 @@

    "postject": ["postject@1.0.0-alpha.6", "", { "dependencies": { "commander": "^9.4.0" }, "bin": { "postject": "dist/cli.js" } }, "sha512-b9Eb8h2eVqNE8edvKdwqkrY6O7kAwmI8kcnBv1NScolYJbo59XUF0noFq+lxbC1yN20bmC0WBEbDC5H/7ASb0A=="],

+    "powershell-utils": ["powershell-utils@0.2.1", "", {}, "sha512-C+y9x90UElAddDZmV4qOx9W53B61PO7cIqWz2dQsWlwswuq4mr8NEwytdGKboYbQlGZ3awrkTeNvcZiZNHnQ8A=="],
+
    "preact": ["preact@11.0.0-beta.0", "", {}, "sha512-IcODoASASYwJ9kxz7+MJeiJhvLriwSb4y4mHIyxdgaRZp6kPUud7xytrk/6GZw8U3y6EFJaRb5wi9SrEK+8+lg=="],

    "preact-render-to-string": ["preact-render-to-string@6.6.5", "", { "peerDependencies": { "preact": ">=10 || >= 11.0.0-0" } }, "sha512-O6MHzYNIKYaiSX3bOw0gGZfEbOmlIDtDfWwN1JJdc/T3ihzRT6tGGSEWE088dWrEDGa1u7101q+6fzQnO9XCPA=="],
@@ -5814,7 +5902,7 @@

    "ws": ["ws@8.21.0", "", { "peerDependencies": { "bufferutil": "^4.0.1", "utf-8-validate": ">=5.0.2" }, "optionalPeers": ["bufferutil", "utf-8-validate"] }, "sha512-Vsp28b7DRcimFQvrqu2Wek3z1iYxDCWqHYB8Qsnk/S4RfaCQzPGPyBNuVjJV3cd6UiKtUtp6sNM77gWvzcCH+g=="],

-    "wsl-utils": ["wsl-utils@0.1.0", "", { "dependencies": { "is-wsl": "^3.1.0" } }, "sha512-h3Fbisa2nKGPxCpm89Hk33lBLsnaGBvctQopaBSOW/uIs6FTe1ATyAnKFJrzVs9vpGdsTe73WF3V4lIsk4Gacw=="],
+    "wsl-utils": ["wsl-utils@1.0.0", "", { "dependencies": { "is-wsl": "^3.1.0", "powershell-utils": "^0.1.0" } }, "sha512-Hl0ZOAs672vg+06kfujwRhoS6/jehvULrlFkuF2dRu6pHgA8U06h3xqNIqNNU1LTXPcedxByAR4GS6pwQK0mgA=="],

    "xdg-basedir": ["xdg-basedir@5.1.0", "", {}, "sha512-GCPAHLvrIH13+c0SuacwvRYj2SxJXQ4kaVTT5xgL3kPrz56XxkF21IGhjSE1+W0aw7gpBWRGXLCPnPby6lSpmQ=="],

@@ -5910,8 +5998,6 @@

    "@astrojs/telemetry/ci-info": ["ci-info@4.4.0", "", {}, "sha512-77PSwercCZU2Fc4sX94eF8k8Pxte6JAwL4/ICZLFjJLqegs7kCuAsqqj/70NQF6TvDpgFjkubQB2FW2ZZddvQg=="],

-    "@astrojs/telemetry/is-docker": ["is-docker@4.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-LHE+wROyG/Y/0ZnbktRCoTix2c1RhgWaZraMZ8o1Q7zCh0VSrICJQO5oqIIISrcSBtrXv0o233w1IYwsWCjTzA=="],
-
    "@aws-crypto/crc32/@aws-sdk/types": ["@aws-sdk/types@3.974.4", "", { "dependencies": { "@smithy/types": "^4.16.1", "tslib": "^2.6.2" } }, "sha512-dSFDNG00MEz0/xl5gxL62giLd1iYyJsTxZ1I1DOj6lC+bbgLB4TRsYClJg3b62dhXT1uATzsTNXPnC+33EJV3A=="],

    "@aws-crypto/crc32c/@aws-sdk/types": ["@aws-sdk/types@3.974.4", "", { "dependencies": { "@smithy/types": "^4.16.1", "tslib": "^2.6.2" } }, "sha512-dSFDNG00MEz0/xl5gxL62giLd1iYyJsTxZ1I1DOj6lC+bbgLB4TRsYClJg3b62dhXT1uATzsTNXPnC+33EJV3A=="],
@@ -6170,6 +6256,8 @@

    "@openauthjs/openauth/jose": ["jose@5.9.6", "", {}, "sha512-AMlnetc9+CV9asI19zHmrgS/WYsWUwCn2R7RzlbJWD7F9eWYUTGyBmU9o6PxngtLGOiDGPRu+Uc4fhKzbpteZQ=="],

+    "@opencode/browser-extension/tailwindcss": ["tailwindcss@4.1.11", "", {}, "sha512-2E9TBm6MDD/xKYe+dvJZAmg3yxIEDNRc0jwlNyDg/4Fil2QcSLjFKGVff0lAf1jjeaArlG/M75Ey/EYr/OJtBA=="],
+
    "@opencode/cli/vite": ["vite@7.3.6", "", { "dependencies": { "esbuild": "^0.27.0 || ^0.28.0", "fdir": "^6.5.0", "picomatch": "^4.0.3", "postcss": "^8.5.6", "rollup": "^4.43.0", "tinyglobby": "^0.2.15" }, "optionalDependencies": { "fsevents": "~2.3.3" }, "peerDependencies": { "@types/node": "^20.19.0 || >=22.12.0", "jiti": ">=1.21.0", "less": "^4.0.0", "lightningcss": "^1.21.0", "sass": "^1.70.0", "sass-embedded": "^1.70.0", "stylus": ">=0.54.8", "sugarss": "^5.0.0", "terser": "^5.16.0", "tsx": "^4.8.1", "yaml": "^2.4.2" }, "optionalPeers": ["@types/node", "jiti", "less", "lightningcss", "sass", "sass-embedded", "stylus", "sugarss", "terser", "tsx", "yaml"], "bin": { "vite": "bin/vite.js" } }, "sha512-4XP60spRGjSZFf1qYH+dJIkK2znL3zQfl9KkOV9MkkRR/3Dls0dxaBsQPTloEc5BLXWPL9vsOxopxyKoMmDueg=="],

    "@opencode/cli/vite-plugin-solid": ["vite-plugin-solid@2.11.10", "", { "dependencies": { "@babel/core": "^7.23.3", "@types/babel__core": "^7.20.4", "babel-preset-solid": "^1.8.4", "merge-anything": "^5.1.7", "solid-refresh": "^0.6.3", "vitefu": "^1.0.4" }, "peerDependencies": { "@testing-library/jest-dom": "^5.16.6 || ^5.17.0 || ^6.*", "solid-js": "^1.7.2", "vite": "^3.0.0 || ^4.0.0 || ^5.0.0 || ^6.0.0 || ^7.0.0" }, "optionalPeers": ["@testing-library/jest-dom"] }, "sha512-Yr1dQybmtDtDAHkii6hXuc1oVH9CPcS/Zb2jN/P36qqcrkNnVPsMTzQ06jyzFPFjj3U1IYKMVt/9ZqcwGCEbjw=="],
@@ -6194,6 +6282,8 @@

    "@opencode/files/wrangler": ["wrangler@4.110.0", "", { "dependencies": { "@cloudflare/kv-asset-handler": "0.5.0", "@cloudflare/unenv-preset": "2.16.1", "blake3-wasm": "2.1.5", "esbuild": "0.28.1", "miniflare": "4.20260708.1", "path-to-regexp": "6.3.0", "unenv": "2.0.0-rc.24", "workerd": "1.20260708.1" }, "optionalDependencies": { "fsevents": "2.3.3" }, "peerDependencies": { "@cloudflare/workers-types": "^5.20260708.1" }, "optionalPeers": ["@cloudflare/workers-types"], "bin": { "wrangler": "bin/wrangler.js", "wrangler2": "bin/wrangler.js", "cf-wrangler": "bin/cf-wrangler.js" } }, "sha512-xZeXKYi7hxQRF5anL+v77RkufJNpF9f3Eqeyqq2QBsETpLZgh0Agj0jJ6JPtkbgn6ukZdh8OK5egsGPWIditgg=="],

+    "@opencode/gui-extensions/vite": ["vite@7.3.6", "", { "dependencies": { "esbuild": "^0.27.0 || ^0.28.0", "fdir": "^6.5.0", "picomatch": "^4.0.3", "postcss": "^8.5.6", "rollup": "^4.43.0", "tinyglobby": "^0.2.15" }, "optionalDependencies": { "fsevents": "~2.3.3" }, "peerDependencies": { "@types/node": "^20.19.0 || >=22.12.0", "jiti": ">=1.21.0", "less": "^4.0.0", "lightningcss": "^1.21.0", "sass": "^1.70.0", "sass-embedded": "^1.70.0", "stylus": ">=0.54.8", "sugarss": "^5.0.0", "terser": "^5.16.0", "tsx": "^4.8.1", "yaml": "^2.4.2" }, "optionalPeers": ["@types/node", "jiti", "less", "lightningcss", "sass", "sass-embedded", "stylus", "sugarss", "terser", "tsx", "yaml"], "bin": { "vite": "bin/vite.js" } }, "sha512-4XP60spRGjSZFf1qYH+dJIkK2znL3zQfl9KkOV9MkkRR/3Dls0dxaBsQPTloEc5BLXWPL9vsOxopxyKoMmDueg=="],
+
    "@opencode/posts/wrangler": ["wrangler@4.110.0", "", { "dependencies": { "@cloudflare/kv-asset-handler": "0.5.0", "@cloudflare/unenv-preset": "2.16.1", "blake3-wasm": "2.1.5", "esbuild": "0.28.1", "miniflare": "4.20260708.1", "path-to-regexp": "6.3.0", "unenv": "2.0.0-rc.24", "workerd": "1.20260708.1" }, "optionalDependencies": { "fsevents": "2.3.3" }, "peerDependencies": { "@cloudflare/workers-types": "^5.20260708.1" }, "optionalPeers": ["@cloudflare/workers-types"], "bin": { "wrangler": "bin/wrangler.js", "wrangler2": "bin/wrangler.js", "cf-wrangler": "bin/cf-wrangler.js" } }, "sha512-xZeXKYi7hxQRF5anL+v77RkufJNpF9f3Eqeyqq2QBsETpLZgh0Agj0jJ6JPtkbgn6ukZdh8OK5egsGPWIditgg=="],

    "@opencode/session-ui/vite": ["vite@7.3.6", "", { "dependencies": { "esbuild": "^0.27.0 || ^0.28.0", "fdir": "^6.5.0", "picomatch": "^4.0.3", "postcss": "^8.5.6", "rollup": "^4.43.0", "tinyglobby": "^0.2.15" }, "optionalDependencies": { "fsevents": "~2.3.3" }, "peerDependencies": { "@types/node": "^20.19.0 || >=22.12.0", "jiti": ">=1.21.0", "less": "^4.0.0", "lightningcss": "^1.21.0", "sass": "^1.70.0", "sass-embedded": "^1.70.0", "stylus": ">=0.54.8", "sugarss": "^5.0.0", "terser": "^5.16.0", "tsx": "^4.8.1", "yaml": "^2.4.2" }, "optionalPeers": ["@types/node", "jiti", "less", "lightningcss", "sass", "sass-embedded", "stylus", "sugarss", "terser", "tsx", "yaml"], "bin": { "vite": "bin/vite.js" } }, "sha512-4XP60spRGjSZFf1qYH+dJIkK2znL3zQfl9KkOV9MkkRR/3Dls0dxaBsQPTloEc5BLXWPL9vsOxopxyKoMmDueg=="],
@@ -6254,11 +6344,7 @@

    "@parcel/watcher/detect-libc": ["detect-libc@1.0.3", "", { "bin": { "detect-libc": "./bin/detect-libc.js" } }, "sha512-pGjwhsmsp4kL2RTz08wcOlGN83otlqHeD/Z5T8GXZB+/YcpQ/dgo+lbU8ZsGxV0HIvqqxo9l7mqYwyYMD9bKDg=="],

-    "@pierre/diffs/diff": ["diff@8.0.3", "", {}, "sha512-qejHi7bcSD4hQAZE0tNAawRK1ZtafHDmMTMkrrIGgSLl7hTnQHmKCeB45xAcbfTqK2zowkM3j3bHt/4b/ARbYQ=="],
-
-    "@pierre/diffs/react": ["react@19.2.8", "", {}, "sha512-PWaYA1L/q9u2u7xYQi+Y3L3Yfnie7XyLeaJICV1MGD6LprsBxcAqGjYyr0eY3p+QdsA+x/Irkt4Qif8D63+Sbw=="],
-
-    "@pierre/diffs/react-dom": ["react-dom@19.2.8", "", { "dependencies": { "scheduler": "^0.27.0" }, "peerDependencies": { "react": "^19.2.8" } }, "sha512-rVprimfGBG3DR+Tq0IQG2DT5PxKth1WIGDmj5yPmlzr4YBe7uyE+Du4oVqTDXZSHGGGXRtTJEGSSePyQCMBglQ=="],
+    "@pierre/diffs/diff": ["diff@9.0.0", "", {}, "sha512-svtcdpS8CgJyqAjEQIXdb3OjhFVVYjzGAPO8WGCmRbrml64SPw/jJD4GoE98aR7r25A0XcgrK3F02yw9R/vhQw=="],

    "@pierre/trees/react": ["react@19.2.8", "", {}, "sha512-PWaYA1L/q9u2u7xYQi+Y3L3Yfnie7XyLeaJICV1MGD6LprsBxcAqGjYyr0eY3p+QdsA+x/Irkt4Qif8D63+Sbw=="],

@@ -6394,8 +6480,6 @@

    "builder-util/js-yaml": ["js-yaml@4.3.1", "", { "dependencies": { "argparse": "^2.0.1" }, "bin": { "js-yaml": "bin/js-yaml.js" } }, "sha512-CY6crGq313MX8GkwvB7tzgp99vjQxY1++5y10/BKN/GUfHqWaOGQMNZkBvqSzsZKWk/ijwHlWzzkLulsGHhjWQ=="],

-    "chrome-launcher/is-wsl": ["is-wsl@2.2.0", "", { "dependencies": { "is-docker": "^2.0.0" } }, "sha512-fKzAra0rGJUUBwGBgNkHZuToZcn+TtXHpeCgmkMJMMYx1sQDYaCSyjJBSCa2nH1DGm7s3n1oBnohoVTBaN7Lww=="],
-
    "chromium-bidi/zod": ["zod@3.25.76", "", {}, "sha512-gzUt/qt81nXsFGKIFcC3YnfEAx5NkunCfnDlvuBSSFS02bcXu4Lmea0AFIUwbLWxWPx3d9p8S5QoaujKcNQxcQ=="],

    "clean-css/source-map": ["source-map@0.6.1", "", {}, "sha512-UjgapumWlbMhkBgzT7Ykc5YXUT46F0iKu8SGXq0bcwP5dz/h0Plj6enJqjz1Zbq2l5WaqYnrVbwWOWMyF3F47g=="],
@@ -6500,6 +6584,10 @@

    "import-in-the-middle/es-module-lexer": ["es-module-lexer@2.3.2", "", {}, "sha512-poHGpORABojJJucnV9KbOavETW8lBVnphkW77ER5/BQ5Fz7oXSoCNek7IH3vR5nRjdsEz926ibFYX8KtLQmdyw=="],

+    "is-inside-container/is-docker": ["is-docker@3.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-eljcgEDlEns/7AXFosB5K/2nCM4P7FQPkGc/DWLy5rmFEWvZayGrik1d9/QIY5nJ4f9YsVvBkA6kJpHn9rISdQ=="],
+
+    "is-wsl/is-docker": ["is-docker@2.2.1", "", { "bin": { "is-docker": "cli.js" } }, "sha512-F+i2BKsFrH66iaUFc0woD8sLy8getkwTwtOBjvs56Cx4CgJDeKQeqfz8wAYiSb8JOprWhHH5p77PbmYCvvUuXQ=="],
+
    "js-beautify/glob": ["glob@10.5.0", "", { "dependencies": { "foreground-child": "^3.1.0", "jackspeak": "^3.1.2", "minimatch": "^9.0.4", "minipass": "^7.1.2", "package-json-from-dist": "^1.0.0", "path-scurry": "^1.11.1" }, "bin": { "glob": "dist/esm/bin.mjs" } }, "sha512-DfXN8DfhJ7NH3Oe7cFmu3NCu1wKbkReJ8TorzSAFbSKrlNaQSKfIzqYqVY8zlbs2NLBbWpRiU52GX2PbaBVNkg=="],

    "js-beautify/nopt": ["nopt@7.2.1", "", { "dependencies": { "abbrev": "^2.0.0" }, "bin": { "nopt": "bin/nopt.js" } }, "sha512-taM24ViiimT/XntxbPyJQzCG+p4EKOpgD3mxFwW38mGjVUrfERQOeY4EDHjdnptttfHuHQXFx+lTP08Q+mLa/w=="],
@@ -6510,8 +6598,6 @@

    "lighthouse/devtools-protocol": ["devtools-protocol@0.0.1663043", "", {}, "sha512-33aOY3ZnBP1dgZsshgaL+/XlsQleiFZgyUaDtdZkEa1nbZhVY1MoDeWjk+wxg25fU924l1ZJfoGNmjjeA/5s1w=="],

-    "lighthouse/open": ["open@8.4.2", "", { "dependencies": { "define-lazy-prop": "^2.0.0", "is-docker": "^2.1.1", "is-wsl": "^2.2.0" } }, "sha512-7x81NCL719oNbsq/3mh+hVrAWmFuEYUqrq/Iw3kUzH8ReypT9QQ0BLoJS7/G9k6N81XjW4qHWtjWwe/9eLy1EQ=="],
-
    "lighthouse/ws": ["ws@7.5.13", "", { "peerDependencies": { "bufferutil": "^4.0.1", "utf-8-validate": "^5.0.2" }, "optionalPeers": ["bufferutil", "utf-8-validate"] }, "sha512-rsKI6xDBFVf4r/x8XyChGK04QR/XHroxs/jUcoWvtEZM8TPU/X/uIY9B1CsSzYws9ZJb/6bbBu7dPhFW00CAoA=="],

    "md-to-react-email/marked": ["marked@7.0.4", "", { "bin": { "marked": "bin/marked.js" } }, "sha512-t8eP0dXRJMtMvBojtkcsA7n48BkauktUKzfkPSCq85ZMTJ0v76Rke4DYz01omYpPTUh4p/f7HePgRo3ebG8+QQ=="],
@@ -6614,8 +6700,6 @@

    "sst/jose": ["jose@5.2.3", "", {}, "sha512-KUXdbctm1uHVL8BYhnyHkgp3zDX5KW8ZhAKVFEfUbU2P8Alpzjb+48hHvjOdQIyPshoblhzsuqOwEEAbtHVirA=="],

-    "storybook/open": ["open@10.2.0", "", { "dependencies": { "default-browser": "^5.2.1", "define-lazy-prop": "^3.0.0", "is-inside-container": "^1.0.0", "wsl-utils": "^0.1.0" } }, "sha512-YgBpdJHPyQ2UE5x+hlSXcnejzAvD0b22U2OuAP+8OnlJT+PjWPxtgmGqKKc+RgTM63U9gN0YzrYc71R2WT/hTA=="],
-
    "storybook-solidjs-vite/semver": ["semver@7.8.1", "", { "bin": { "semver": "bin/semver.js" } }, "sha512-rkVq3IXh+4FDGch+KwzX3aV9W3kO54GyEgpvBzSyctDA6Xtd7RJQV1xmXbeQp5v7+VzLOfVqiutSE6GICgPFvg=="],

    "storybook-solidjs-vite/vite": ["vite@7.1.11", "", { "dependencies": { "esbuild": "^0.25.0", "fdir": "^6.5.0", "picomatch": "^4.0.3", "postcss": "^8.5.6", "rollup": "^4.43.0", "tinyglobby": "^0.2.15" }, "optionalDependencies": { "fsevents": "~2.3.3" }, "peerDependencies": { "@types/node": "^20.19.0 || >=22.12.0", "jiti": ">=1.21.0", "less": "^4.0.0", "lightningcss": "^1.21.0", "sass": "^1.70.0", "sass-embedded": "^1.70.0", "stylus": ">=0.54.8", "sugarss": "^5.0.0", "terser": "^5.16.0", "tsx": "^4.8.1", "yaml": "^2.4.2" }, "optionalPeers": ["@types/node", "jiti", "less", "lightningcss", "sass", "sass-embedded", "stylus", "sugarss", "terser", "tsx", "yaml"], "bin": { "vite": "bin/vite.js" } }, "sha512-uzcxnSDVjAopEUjljkWh8EIrg6tlzrjFUfMcR1EVsRDGwf/ccef0qQPRyOrROwhrTDaApueq+ja+KLPlzR/zdg=="],
@@ -6702,6 +6786,10 @@

    "write-file-atomic/signal-exit": ["signal-exit@4.1.0", "", {}, "sha512-bzyZ1e88w9O1iNJbKnOlvYTrWPDl46O1bG0D3XInv+9tkPrxrN8jUUTiFlDkkmKWgn1M6CfIA13SuGqOa9Korw=="],

+    "wsl-utils/is-wsl": ["is-wsl@3.1.1", "", { "dependencies": { "is-inside-container": "^1.0.0" } }, "sha512-e6rvdUCiQCAuumZslxRJWR/Doq4VpPR82kqclvcS0efgt430SlGIk05vdCN58+VrzgtIcfNODjozVielycD4Sw=="],
+
+    "wsl-utils/powershell-utils": ["powershell-utils@0.1.0", "", {}, "sha512-dM0jVuXJPsDN6DvRpea484tCUaMiXWjuCn++HGTqUWzGDjv5tZkEZldAJ/UMlqRYGFrD/etByo4/xOuC/snX2A=="],
+
    "yaml-language-server/prettier": ["prettier@3.9.6", "", { "bin": { "prettier": "bin/prettier.cjs" } }, "sha512-OpN0zzVdiaiAhxpuuj5efpIS4sY9j7bY6uR5mnj5yPzGkdkjNKSJeUThPb60Jw29QuAZgA4o+/iB49kFiaBX6g=="],

    "yaml-language-server/request-light": ["request-light@0.5.8", "", {}, "sha512-3Zjgh+8b5fhRJBQZoy+zbVKpAQGLyka0MPgW3zruTF4dFFJ8Fqcfu9YsAvi/rvdcaTeWG3MkbZv4WKxAn/84Lg=="],
@@ -7154,8 +7242,6 @@

    "@oxc-resolver/binding-wasm32-wasi/@emnapi/core/@emnapi/wasi-threads": ["@emnapi/wasi-threads@1.2.2", "", { "dependencies": { "tslib": "^2.4.0" } }, "sha512-c95qOXkHdydNKhscBTebqEC1CVAZpyqOfVfBzQ1qgzyl3gfeldUjIggDbIZgDKsHLgnsM+igH7TJ/eAasaVuMA=="],

-    "@pierre/diffs/react-dom/scheduler": ["scheduler@0.27.0", "", {}, "sha512-eNv+WrVbKu1f3vbYJT/xtiF5syA5HPIMtf9IgY/nKg0sWqzAUEvqY/xm7OcZc/qafLx/iO9FgOmeSAp4v5ti/Q=="],
-
    "@pierre/trees/react-dom/scheduler": ["scheduler@0.27.0", "", {}, "sha512-eNv+WrVbKu1f3vbYJT/xtiF5syA5HPIMtf9IgY/nKg0sWqzAUEvqY/xm7OcZc/qafLx/iO9FgOmeSAp4v5ti/Q=="],

    "@puppeteer/browsers/yargs/cliui": ["cliui@9.0.1", "", { "dependencies": { "string-width": "^7.2.0", "strip-ansi": "^7.1.0", "wrap-ansi": "^9.0.0" } }, "sha512-k7ndgKhwoQveBL+/1tqGJYNz097I7WOvwbmmU2AR5+magtbjPWQTS1C5vzGkBC8Ym8UWRzfKUzUUqFLypY4Q+w=="],
@@ -7336,8 +7422,6 @@

    "builder-util/js-yaml/argparse": ["argparse@2.0.1", "", {}, "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q=="],

-    "chrome-launcher/is-wsl/is-docker": ["is-docker@2.2.1", "", { "bin": { "is-docker": "cli.js" } }, "sha512-F+i2BKsFrH66iaUFc0woD8sLy8getkwTwtOBjvs56Cx4CgJDeKQeqfz8wAYiSb8JOprWhHH5p77PbmYCvvUuXQ=="],
-
    "cliui/string-width/emoji-regex": ["emoji-regex@8.0.0", "", {}, "sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A=="],

    "cliui/strip-ansi/ansi-regex": ["ansi-regex@5.0.1", "", {}, "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ=="],
@@ -7398,12 +7482,6 @@

    "lazystream/readable-stream/string_decoder": ["string_decoder@1.1.1", "", { "dependencies": { "safe-buffer": "~5.1.0" } }, "sha512-n/ShnvDi6FHbbVfviro+WojiFzv+s8MPMHBczVePfUpDJLwoLT0ht1l4YwBCbi8pJAveEEdnkHyPyTP/mzRfwg=="],

-    "lighthouse/open/define-lazy-prop": ["define-lazy-prop@2.0.0", "", {}, "sha512-Ds09qNh8yw3khSjiJjiUInaGX9xlqZDY7JVryGxdxV7NPeuqQfplOpQ66yJFZut3jLa5zOwkXw1g9EI2uKh4Og=="],
-
-    "lighthouse/open/is-docker": ["is-docker@2.2.1", "", { "bin": { "is-docker": "cli.js" } }, "sha512-F+i2BKsFrH66iaUFc0woD8sLy8getkwTwtOBjvs56Cx4CgJDeKQeqfz8wAYiSb8JOprWhHH5p77PbmYCvvUuXQ=="],
-
-    "lighthouse/open/is-wsl": ["is-wsl@2.2.0", "", { "dependencies": { "is-docker": "^2.0.0" } }, "sha512-fKzAra0rGJUUBwGBgNkHZuToZcn+TtXHpeCgmkMJMMYx1sQDYaCSyjJBSCa2nH1DGm7s3n1oBnohoVTBaN7Lww=="],
-
    "miniflare/sharp/@img/sharp-darwin-arm64": ["@img/sharp-darwin-arm64@0.33.5", "", { "optionalDependencies": { "@img/sharp-libvips-darwin-arm64": "1.0.4" }, "os": "darwin", "cpu": "arm64" }, "sha512-UT4p+iz/2H4twwAoLCqfA9UH5pI6DggwKEGuaPy7nCVQ8ZsiY5PIcrRvD1DzuY3qYL07NtIQcWnBSY/heikIFQ=="],

    "miniflare/sharp/@img/sharp-darwin-x64": ["@img/sharp-darwin-x64@0.33.5", "", { "optionalDependencies": { "@img/sharp-libvips-darwin-x64": "1.0.4" }, "os": "darwin", "cpu": "x64" }, "sha512-fyHac4jIc1ANYGRDxtiqelIbdWkIuQaI84Mv45KvGRRxSAa7o7d1ZKAOBaYbnepLC1WqxfpimdeWfvqqSGwR2Q=="],
@@ -7674,6 +7752,10 @@

    "@astrojs/starlight/@astrojs/mdx/@astrojs/markdown-remark/shiki": ["shiki@3.23.0", "", { "dependencies": { "@shikijs/core": "3.23.0", "@shikijs/engine-javascript": "3.23.0", "@shikijs/engine-oniguruma": "3.23.0", "@shikijs/langs": "3.23.0", "@shikijs/themes": "3.23.0", "@shikijs/types": "3.23.0", "@shikijs/vscode-textmate": "^10.0.2", "@types/hast": "^3.0.4" } }, "sha512-55Dj73uq9ZXL5zyeRPzHQsK7Nbyt6Y10k5s7OjuFZGMhpp4r/rsLBH0o/0fstIzX1Lep9VxefWljK/SKCzygIA=="],

+    "@astrojs/starlight/astro/@astrojs/telemetry/is-docker": ["is-docker@3.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-eljcgEDlEns/7AXFosB5K/2nCM4P7FQPkGc/DWLy5rmFEWvZayGrik1d9/QIY5nJ4f9YsVvBkA6kJpHn9rISdQ=="],
+
+    "@astrojs/starlight/astro/@astrojs/telemetry/is-wsl": ["is-wsl@3.1.1", "", { "dependencies": { "is-inside-container": "^1.0.0" } }, "sha512-e6rvdUCiQCAuumZslxRJWR/Doq4VpPR82kqclvcS0efgt430SlGIk05vdCN58+VrzgtIcfNODjozVielycD4Sw=="],
+
    "@astrojs/starlight/astro/p-queue/p-timeout": ["p-timeout@6.1.4", "", {}, "sha512-MyIV3ZA/PmyBN/ud8vV9XzwTrNtR4jFrObymZYnZqMmW0zA8Z17vnT0rBgFE/TlohB+YCHqXMgZzb3Csp49vqg=="],

    "@astrojs/starlight/astro/sharp/@img/sharp-darwin-arm64": ["@img/sharp-darwin-arm64@0.33.5", "", { "optionalDependencies": { "@img/sharp-libvips-darwin-arm64": "1.0.4" }, "os": "darwin", "cpu": "arm64" }, "sha512-UT4p+iz/2H4twwAoLCqfA9UH5pI6DggwKEGuaPy7nCVQ8ZsiY5PIcrRvD1DzuY3qYL07NtIQcWnBSY/heikIFQ=="],
@@ -8076,6 +8158,10 @@

    "@opencode/web/@astrojs/cloudflare/wrangler/workerd": ["workerd@1.20260708.1", "", { "optionalDependencies": { "@cloudflare/workerd-darwin-64": "1.20260708.1", "@cloudflare/workerd-darwin-arm64": "1.20260708.1", "@cloudflare/workerd-linux-64": "1.20260708.1", "@cloudflare/workerd-linux-arm64": "1.20260708.1", "@cloudflare/workerd-windows-64": "1.20260708.1" }, "bin": { "workerd": "bin/workerd" } }, "sha512-WAK+Kt/VVCSldH2qSr8lx46XCJ4Q+bdlHNaFqUtOHthBEIB8C1N8HVW+VOLrxDoTCk0NGNv0zajnBeQK4JOB9w=="],

+    "@opencode/web/astro/@astrojs/telemetry/is-docker": ["is-docker@3.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-eljcgEDlEns/7AXFosB5K/2nCM4P7FQPkGc/DWLy5rmFEWvZayGrik1d9/QIY5nJ4f9YsVvBkA6kJpHn9rISdQ=="],
+
+    "@opencode/web/astro/@astrojs/telemetry/is-wsl": ["is-wsl@3.1.1", "", { "dependencies": { "is-inside-container": "^1.0.0" } }, "sha512-e6rvdUCiQCAuumZslxRJWR/Doq4VpPR82kqclvcS0efgt430SlGIk05vdCN58+VrzgtIcfNODjozVielycD4Sw=="],
+
    "@opencode/web/astro/js-yaml/argparse": ["argparse@2.0.1", "", {}, "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q=="],

    "@opencode/web/astro/p-queue/p-timeout": ["p-timeout@6.1.4", "", {}, "sha512-MyIV3ZA/PmyBN/ud8vV9XzwTrNtR4jFrObymZYnZqMmW0zA8Z17vnT0rBgFE/TlohB+YCHqXMgZzb3Csp49vqg=="],
@@ -8216,6 +8302,10 @@

    "archiver-utils/glob/path-scurry/lru-cache": ["lru-cache@10.4.3", "", {}, "sha512-JNAzZcXrCt42VGLuYz0zfAzDfAvJWW6AfYlDBQyDV5DClI2m5sAmK+OIO7s59XfsRsWHp02jAJrRadPRGTt6SQ=="],

+    "astro-expressive-code/astro/@astrojs/telemetry/is-docker": ["is-docker@3.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-eljcgEDlEns/7AXFosB5K/2nCM4P7FQPkGc/DWLy5rmFEWvZayGrik1d9/QIY5nJ4f9YsVvBkA6kJpHn9rISdQ=="],
+
+    "astro-expressive-code/astro/@astrojs/telemetry/is-wsl": ["is-wsl@3.1.1", "", { "dependencies": { "is-inside-container": "^1.0.0" } }, "sha512-e6rvdUCiQCAuumZslxRJWR/Doq4VpPR82kqclvcS0efgt430SlGIk05vdCN58+VrzgtIcfNODjozVielycD4Sw=="],
+
    "astro-expressive-code/astro/js-yaml/argparse": ["argparse@2.0.1", "", {}, "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q=="],

    "astro-expressive-code/astro/p-queue/p-timeout": ["p-timeout@6.1.4", "", {}, "sha512-MyIV3ZA/PmyBN/ud8vV9XzwTrNtR4jFrObymZYnZqMmW0zA8Z17vnT0rBgFE/TlohB+YCHqXMgZzb3Csp49vqg=="],
@@ -8318,6 +8408,10 @@

    "temp/rimraf/glob/minimatch": ["minimatch@3.1.5", "", { "dependencies": { "brace-expansion": "^1.1.7" } }, "sha512-VgjWUsnnT6n+NUk6eZq77zeFdpW2LWDzP6zFGrCbHXiYNul5Dzqk2HHQ5uFH2DNW5Xbp8+jVzaeNt94ssEEl4w=="],

+    "toolbeam-docs-theme/astro/@astrojs/telemetry/is-docker": ["is-docker@3.0.0", "", { "bin": { "is-docker": "cli.js" } }, "sha512-eljcgEDlEns/7AXFosB5K/2nCM4P7FQPkGc/DWLy5rmFEWvZayGrik1d9/QIY5nJ4f9YsVvBkA6kJpHn9rISdQ=="],
+
+    "toolbeam-docs-theme/astro/@astrojs/telemetry/is-wsl": ["is-wsl@3.1.1", "", { "dependencies": { "is-inside-container": "^1.0.0" } }, "sha512-e6rvdUCiQCAuumZslxRJWR/Doq4VpPR82kqclvcS0efgt430SlGIk05vdCN58+VrzgtIcfNODjozVielycD4Sw=="],
+
    "toolbeam-docs-theme/astro/js-yaml/argparse": ["argparse@2.0.1", "", {}, "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q=="],

    "toolbeam-docs-theme/astro/p-queue/p-timeout": ["p-timeout@6.1.4", "", {}, "sha512-MyIV3ZA/PmyBN/ud8vV9XzwTrNtR4jFrObymZYnZqMmW0zA8Z17vnT0rBgFE/TlohB+YCHqXMgZzb3Csp49vqg=="],
@@ -2,11 +2,11 @@
  "nodes": {
    "nixpkgs": {
      "locked": {
-        "lastModified": 1776683584,
-        "narHash": "sha256-NuTLMrr10Tng72hurYG8jYQ4XKK8wnpJmOGcPiis96g=",
+        "lastModified": 1790510107,
+        "narHash": "sha256-EVMNYv7hYDDD9TGVT/hIyTYgpiXA8y3m5xIEIxuGNU0=",
        "owner": "NixOS",
        "repo": "nixpkgs",
-        "rev": "9dd5558b06dbdacbf635a3dd36dce1b1a7ee3a89",
+        "rev": "3181085bfd08663b6b9e60bc7a8395c2aaa741bd",
        "type": "github"
      },
      "original": {
@@ -12,7 +12,6 @@
        "aarch64-linux"
        "x86_64-linux"
        "aarch64-darwin"
-        "x86_64-darwin"
      ];
      forEachSystem = f: nixpkgs.lib.genAttrs systems (system: f nixpkgs.legacyPackages.${system});
      rev = self.shortRev or self.dirtyShortRev or "dirty";
@@ -22,7 +21,7 @@
        default = pkgs.mkShell {
          packages = with pkgs; [
            bun
-            nodejs_20
+            nodejs
            pkg-config
            openssl
            git
@@ -13,7 +13,12 @@
  opencode,
 }:
 let
-  electron = callPackage ./electron.nix { };
+  electronPin =
+    (lib.pipe ../packages/desktop/package.json [
+      builtins.readFile
+      builtins.fromJSON
+    ]).devDependencies.electron;
+  electron = callPackage ./electron.nix { inherit electronPin; };
 in
 stdenv.mkDerivation (finalAttrs: {
  pname = "opencode-desktop";
@@ -66,7 +71,7 @@ stdenv.mkDerivation (finalAttrs: {
    ''
    # https://github.com/electron/electron/issues/31121
    # mac builds use a .app bundle which doesnt have this issue
-    + lib.optionalString stdenv.isLinux ''
+    + lib.optionalString stdenv.hostPlatform.isLinux ''
      substituteInPlace \
        packages/desktop/src/main/windows/appearance.ts \
        packages/desktop/src/main/service/desktop-cli.ts \
@@ -74,6 +79,7 @@ stdenv.mkDerivation (finalAttrs: {
    '';

  preBuild = ''
+    echo "electron ${electron.version} from nixpkgs ${lib.version}, package.json pins ${electronPin}"
    cp -r "${electron.dist}" $HOME/.electron-dist
    chmod -R u+w $HOME/.electron-dist

@@ -89,8 +95,15 @@ stdenv.mkDerivation (finalAttrs: {

    export OPENCODE_CLI_DIST="$TMPDIR/desktop-cli"
    cli_package=$(bun -e 'import { getCurrentCli } from "./scripts/utils.ts"; console.log(getCurrentCli().package.replace("@opencode/", ""))')
+    # copyBuiltCliToResources joins this dist with the npm package name getCurrentCli()
+    # reports, not the Nix build's name. It reads only .version from the manifest and
+    # writes it as opencode-cli.version beside the binary.
    mkdir -p "$OPENCODE_CLI_DIST/$cli_package/bin"
    cp ${lib.getExe opencode} "$OPENCODE_CLI_DIST/$cli_package/bin/opencode"
+    # OPENCODE_VERSION is what the bundled CLI prints for --version, so the manifest
+    # and the executable cannot drift.
+    bun -e 'await Bun.write(process.argv[1], JSON.stringify({ version: process.env.OPENCODE_VERSION }) + "\n")' \
+      "$OPENCODE_CLI_DIST/$cli_package/package.json"

    bun run build
    npx electron-builder --dir \
@@ -138,6 +151,13 @@ stdenv.mkDerivation (finalAttrs: {
    "libc.musl-x86_64.so.1"
  ];

+  passthru = {
+    # electronVersion is what ships; electronPin is what packages/desktop/package.json
+    # asks for. They differ whenever nixpkgs carries no release of the pinned minor.
+    electronVersion = electron.version;
+    inherit electronPin;
+  };
+
  meta = {
    description = "OpenCode Desktop App";
    mainProgram = "opencode-desktop";
@@ -1,13 +1,10 @@
-{ callPackage, path }:
+{ lib, pkgs, electronPin }:
 let
-  version = (builtins.fromJSON (builtins.readFile ../packages/desktop/package.json)).devDependencies.electron;
+  # Nixpkgs owns the release hashes, so bumping the pin no longer means editing this repo. Only the
+  # major is delegated, so what gets built trails the pin whenever nixpkgs has not shipped it yet.
+  # That is safe: the bundle ships one native addon, node-pty's Node-API prebuild, and
+  # only the win32 WSL runtime loads it, so nothing in the main process binds the
+  # Electron ABI.
+  major = lib.versions.major electronPin;
 in
-(callPackage (path + "/pkgs/development/tools/electron/binary/generic.nix") { }) version {
-  # Electron 42.10.1 SHASUMS256.txt; update with the desktop package version.
-  aarch64-linux = "20e68d6c4e47f3ebf59de7c6b1f8b8bec6a6ebda6a451132f9b465f3f13ce467";
-  x86_64-linux = "2452b27112d92387471fa2488aafac85d79ea3f2ee1216c0abd5150d6c12362b";
-  aarch64-darwin = "ac7194a3dfd81930ba35355c01620262c1254752859b42dcb8f4b9e4d174a871";
-  x86_64-darwin = "4489aba55477a0082266cb690db1c829503ba3338048599d8fd243953df37dab";
-  # fetchzip hashes the unpacked headers, not the release tarball.
-  headers = "sha256-4eUy3BZVvxTl7KUOsxio7769lL6ag/ecbeK+qLURWMI=";
-}
+pkgs."electron_${major}-bin" or (throw "nixpkgs ${lib.version} carries no prebuilt electron ${major}: run `nix flake update nixpkgs`, or pin a major nixpkgs still carries")
@@ -1,8 +1,7 @@
 {
  "nodeModules": {
-    "x86_64-linux": "sha256-aQQQhaUlAhpfqzH0vNi0IJ1cg7FQHIKYzxeq5d8PZoU=",
-    "aarch64-linux": "sha256-r9aDFu3UYmudmmYPhzCrpFvQlaejXc8V1IzLtG3jZPc=",
-    "aarch64-darwin": "sha256-B0m41LelD7d61vPHGIZZSO/cU7gbjHDJHt6oxNRRM8Q=",
-    "x86_64-darwin": "sha256-9TWJsyI3Y6BMomtGSgqA1th9LpxpqP4F5Tl/GyexVYw="
+    "x86_64-linux": "sha256-EEBz2IQ14YPShAzY13zpZ7btq/EuSw86pZ6nrizc9BI=",
+    "aarch64-linux": "sha256-X7om4U4OlKL4OddEMcyKL+1DCQXc9/f1xzzopE16Zmk=",
+    "aarch64-darwin": "sha256-3bejEuX3AGv4SR/16O8KjNSO2t9WDaWCHRbBXsJ6z8A="
  }
 }
@@ -81,6 +81,5 @@ stdenvNoCC.mkDerivation {
    "aarch64-linux"
    "x86_64-linux"
    "aarch64-darwin"
-    "x86_64-darwin"
  ];
 }
@@ -12,7 +12,7 @@
  installShellFiles,
  versionCheckHook,
  writableTmpDirAsHomeHook,
-  node_modules ? callPackage ./node-modules.nix { },
+  node_modules ? callPackage ./node_modules.nix { },
 }:
 stdenvNoCC.mkDerivation (finalAttrs: {
  pname = "opencode";
@@ -85,14 +85,30 @@ stdenvNoCC.mkDerivation (finalAttrs: {
  '';

  postInstall = lib.optionalString (stdenvNoCC.buildPlatform.canExecute stdenvNoCC.hostPlatform) ''
-    # trick yargs into also generating zsh completions
+    # v2 dropped the `completion` subcommand; --completions is the global flag.
+    # --completions also accepts sh, which emits the same script as bash.
+    # staged to files, substitute below rejects anything that is not a regular file
+    $out/bin/opencode --completions bash > opencode.bash
+    $out/bin/opencode --completions zsh > _opencode
+    $out/bin/opencode --completions fish > opencode.fish
+
    installShellCompletion --cmd opencode \
-      --bash <($out/bin/opencode completion) \
-      --zsh <(SHELL=/bin/zsh $out/bin/opencode completion)
+      --bash opencode.bash \
+      --fish opencode.fish \
+      --zsh _opencode
+
+    # OPENCODE_CLI_NAME is a build-time define, so the opencode2 copies are
+    # renamed rather than regenerated. --replace-fail is a global literal
+    # substitution, so any lowercase opencode that later appears in a
+    # description or help text ships as opencode2 in the opencode2 copy.
+    substitute opencode.bash opencode2.bash --replace-fail opencode opencode2
+    substitute _opencode _opencode2 --replace-fail opencode opencode2
+    substitute opencode.fish opencode2.fish --replace-fail opencode opencode2

    installShellCompletion --cmd opencode2 \
-      --bash <($out/bin/opencode2 completion) \
-      --zsh <(SHELL=/bin/zsh $out/bin/opencode2 completion)
+      --bash opencode2.bash \
+      --fish opencode2.fish \
+      --zsh _opencode2
  '';

  nativeInstallCheckInputs = [
@@ -2,7 +2,7 @@
  "$schema": "https://json.schemastore.org/package.json",
  "name": "opencode",
  "description": "AI-powered development tool",
-  "version": "2.0.16",
+  "version": "2.0.22",
  "private": true,
  "type": "module",
  "packageManager": "bun@1.4.2",
@@ -18,7 +18,7 @@
    "dev:www": "bun run --cwd services/www dev",
    "dev:storybook": "bun --cwd packages/storybook storybook",
    "bench:devex": "bun run --cwd packages/app test:bench:devex",
-    "lint": "oxlint",
+    "lint": "oxlint && ast-grep scan -c script/ast-grep/gui-extensions/sgconfig.yml",
    "lint:effect-patterns": "ast-grep scan -c script/ast-grep/sgconfig.yml packages/util/src packages/core/src packages/server/src packages/protocol/src packages/cli/src",
    "lint:effect-simplifications": "ast-grep scan -c script/ast-grep/effect-simplifications/sgconfig.yml --off=unused-suppression packages",
    "test:lint-rules": "ast-grep test -c script/ast-grep/sgconfig.yml",
@@ -52,9 +52,9 @@
      "@octokit/rest": "22.0.0",
      "@hono/standard-validator": "0.2.0",
      "@hono/zod-validator": "0.4.2",
-      "@opentui/core": "0.5.12",
-      "@opentui/keymap": "0.5.12",
-      "@opentui/solid": "0.5.12",
+      "@opentui/core": "0.5.14",
+      "@opentui/keymap": "0.5.14",
+      "@opentui/solid": "0.5.14",
      "@tanstack/solid-virtual": "3.13.37",
      "@shikijs/stream": "4.4.3",
      "@standard-schema/spec": "1.1.0",
@@ -68,7 +68,7 @@
      "@tsconfig/bun": "1.0.9",
      "@cloudflare/workers-types": "4.20251008.0",
      "@openauthjs/openauth": "0.0.0-20250322224806",
-      "@pierre/diffs": "1.2.10",
+      "@pierre/diffs": "1.5.1",
      "opentui-spinner": "0.0.7",
      "@solid-primitives/event-listener": "2.4.6",
      "@solid-primitives/media": "2.3.6",
@@ -117,10 +117,11 @@
    "@actions/artifact": "5.0.1",
    "@ast-grep/cli": "0.44.0",
    "@opencode/client": "workspace:*",
-    "@types/react": "19.2.17",
-    "@types/react-dom": "19.2.3",
+    "@oxlint/plugins": "1.60.0",
    "@tsconfig/bun": "catalog:",
    "@types/mime-types": "3.0.1",
+    "@types/react": "19.2.17",
+    "@types/react-dom": "19.2.3",
    "@typescript/native-preview": "catalog:",
    "glob": "13.0.5",
    "husky": "9.1.7",
@@ -162,6 +163,7 @@
    "@types/node": "catalog:",
    "bun-types": "1.4.2",
    "effect": "catalog:",
+    "open": "11.0.4",
    "solid-js": "catalog:"
  },
  "patchedDependencies": {
@@ -12,9 +12,9 @@

 Per-type constructors live on the type, not as top-level re-exports. Use `Message.system(...)`, `Message.user(...)`, `Message.assistant(...)`, `Message.tool(...)`, `Message.media(...)`, `LanguageModel.make(...)`, `ToolDefinition.make(...)`, `ToolCallPart.make(...)`, `ToolResultPart.make(...)`, `ToolChoice.make(...)`, `ToolChoice.named(...)`, `SystemPart.make(...)`, and `GenerationOptions.make(...)` directly. The top-level `LLM` namespace is reserved for request-shaped call APIs: `LLM.request`, `LLM.generate`, `LLM.stream`, and `LLM.generateObject`. `LLM.generate`/`LLM.stream` and Promise `ai.llm.generate`/`ai.llm.stream` accept ergonomic input or an `LLMRequest`; both paths use the same canonical request. Core still builds, logs, replays, and updates that durable `LLMRequest` boundary. Use `LLMRequest.update(...)` when deriving canonical request data; do not add a duplicate `LLM.updateRequest(...)` path.

-Modality namespaces mirror `LLM` exactly: `Image.request`, `Image.generate`, `Image.stream` (later `Video`, `Speech`, `Transcription`). Common request fields (`images`, `mask`, `n`, `size`, `aspectRatio`, `seed`, `format`) lower natively or fail with a typed `AIError`; provider-native controls always live under `providerOptions`, never under a modality-specific `options` key.
+Modality namespaces mirror `LLM` exactly: `Image.request`, `Image.generate`, `Image.stream`, and the same for `Video`, `Speech`, and `Transcription`. Common request fields (`images`, `mask`, `n`, `size`, `aspectRatio`, `seed`, `format`) lower natively or fail with a typed `AIError`; provider-native controls always live under `providerOptions`, never under a modality-specific `options` key.

-Media payloads are always `Media.Asset` (`src/media.ts`). Construct them with `Media.bytes`, `Media.base64`, `Media.url`, `Media.ref`, `Media.fromDataUrl`, or `Media.file`; never introduce a parallel `data: string | Uint8Array` shape. `MediaPart.media`, `ImageRequest.images`/`mask`, `ImageResponse.images`, and the `media` `LLMEvent` all share it. Protocols branch on `asset.source.type` and `asset.kind` and use `ProviderShared.inlineMedia` / `requireInlineMedia` / `mediaUrl` / `MediaInput.refID` rather than re-deriving base64 or URL handling.
+Media payloads are always `Media.Asset` (`src/media.ts`). Construct them with `Media.bytes`, `Media.base64`, `Media.url`, `Media.ref`, `Media.fromDataUrl`, or `Media.file`; never introduce a parallel `data: string | Uint8Array` shape. `MediaPart.media`, `ImageRequest.images`/`mask`, `ImageResponse.images`, and the `media` `LLMEvent` all share it. Protocols branch on `asset.source.type` and `asset.kind` and use `ProviderShared.requireInlineMedia` / `inlineRequired` / `mediaUrl` / `mediaReference` and `MediaInput.inlineBytes` / `refID` rather than re-deriving base64 or URL handling.

 `schema/messages.ts → media.ts → route/executor-service.ts` is an accepted runtime dependency from the schema layer on the executor service tag: `Media.Asset.bytes()` must be able to download `url` sources, and the tag lives in that leaf module precisely so the schema barrel never imports the executor implementation (which imports the schema barrel back). Do not move the tag into `route/executor.ts` or import `route/executor.ts` from `src/schema/*` or `src/media.ts`.

@@ -98,11 +98,11 @@ When a provider supports multiple physical transports, selection remains executi

 Media does not fit the SSE-frames-to-event-state-machine LLM route. `MediaRoute.inline(...)` / `queued(...)` / `stream(...)` (`src/route/media.ts`) compose a `MediaProtocol` kind with `Endpoint` and `Auth` and own the transport plumbing: `http` option merging, URL/query rendering, auth headers, JSON vs multipart encoding, and handing the response back to the protocol. `MediaProtocol.inline` (`src/route/media-protocol.ts`) is `body.from(request)` plus `response.decode(response, context)`; each protocol declares `const route = MediaProtocol.identity({ id, name, provider })` once and decodes through `route.decodeJson` / `route.text` / `route.decodeStarted` so decode failures retain the raw body and HTTP context, raising `route.unsupported(operation, message)` for requests it cannot lower, and passes `route` as the first argument to `MediaProtocol.inline` / `queued` / `stream`. `Generation` (`src/generation.ts`) is the provider-neutral handle for a queued generation over a `GenerationRoute` (`status`, `result`, `cancel`). Image protocol files follow the same section order as LLM protocols and declare unsupported common fields once through the protocol's `unsupported` list.

-`MediaProtocol.queued` is the submit-then-poll kind every video route uses: `start` (body + decode into `{ token, snapshot }`), `status`, `result`, and optional `cancel`, each addressed by a route-owned `token` whose `Schema.Codec` makes it serializable. `MediaRoute.inline` and `MediaRoute.queued` compose the two kinds with `Endpoint` and `Auth`; the queued route decodes the token once at the boundary (`start` output or `resume` input) and closes over it in a token-free `GenerationRoute` (`status`/`result`/`cancel` are plain Effects), so `Generation` never sees the token's shape and only carries the encoded JSON for persistence. Polls reuse the route's auth and deployment headers plus the request's `http` overlay after `start`, and resolve relative paths against the route base URL (provider-issued absolute URLs such as fal's `status_url` pass through). `result` is always its own GET even when the provider returns output inside the status document, so `Generation.await` behaves the same after `start` and after `resume`. `PollContext.auth` carries only what `Auth` added or changed so protocols can hand download credentials to output assets as transient `Media.Asset.headers` (Veo) — never part of `source` or JSON. Status strings map through a per-protocol `STATUS` table via `MediaProtocol.status`; terminal generations without output fail through `output.ended` / `output.contentPolicy` with the provider document on `reason.body`. `GenerationAwaitOptions` (`AwaitOptions` in `src/generation.ts`, `{ poll?: Poll }`) is the one options type for `await`, `events`, `Video.generate`, and `Video.stream`.
+`MediaProtocol.queued` is the submit-then-poll kind every video route uses: `start` (body + decode into `{ token, snapshot }`), `status`, `result`, and optional `cancel` (with `activeOnly` when the provider's cancel endpoint deletes finished work, as Runway's does: the route refreshes status first and skips terminal generations), each addressed by a route-owned `token` whose `Schema.Codec` makes it serializable. `MediaRoute.inline` and `MediaRoute.queued` compose the two kinds with `Endpoint` and `Auth`; the queued route decodes the token once at the boundary (`start` output or `resume` input) and closes over it in a token-free `GenerationRoute` (`status`/`result`/`cancel` are plain Effects), so `Generation` never sees the token's shape and only carries the encoded JSON for persistence. Polls reuse the route's auth and deployment headers plus the request's `http` overlay after `start`, and resolve relative paths against the route base URL (provider-issued absolute URLs such as fal's `status_url` pass through). `result` is always its own GET even when the provider returns output inside the status document, so `Generation.await` behaves the same after `start` and after `resume`. `PollContext.auth` carries only what `Auth` added or changed so protocols can hand download credentials to output assets as transient `Media.Asset.headers` (Veo) — never part of `source` or JSON. Status strings map through a per-protocol `STATUS` table via `MediaProtocol.status`; terminal generations without output fail through `output.ended` / `output.contentPolicy` with the provider document on `reason.body`; a `failed` generation maps the provider's error code through a per-protocol `FAILURE` table via `MediaProtocol.failure` so rejected inputs are not reported as retryable `ProviderInternal`. `GenerationAwaitOptions` (`AwaitOptions` in `src/generation.ts`, `{ poll?: Poll }`) is the one options type for `await`, `events`, `Video.generate`, and `Video.stream`.

 `MediaProtocol.stream` is the incremental kind every speech route uses, with the same discipline as LLM protocols. `MediaRoute.stream` submits the caller's request as `MediaProtocol.Addressed<Request>` (`{ ...request, mode }`, `mode: "generate" | "stream"`), so one provider stays one protocol: `body.from`, the endpoint path, and `frames` read `request.mode` to pick the body, path, and framing. `frames(bytes, context)` returns frames — `Framing.sse`, `Framing.lines`, `Framing.document` (a single-document response shaped like a streamed record), or the raw `bytes` for chunked audio. `initial()` is fresh per-response parser state; `step` folds each frame into it and emits modality events; `finish(state, context)` runs once after the last frame with the request, body, and observed `http` (header-only usage lives there) and emits exactly one terminal event or fails with `route.incomplete()`. Keep parser state to real accumulators and derive anything the request or body determines in `finish`. `generate` runs the same stream and folds it with the modality's `collect`. Request-derived URL parameters go on the body's `query` (array values repeat the parameter), applied before route and caller `http.query`. Decode frames with `route.decodeFrame` and raise stream-time failures with `route.frameError` (the frame stays on `reason.body`); protocols never thread HTTP context, because the route fills `reason.http` on stream errors that lack it. Speech protocols share `protocols/utils/speech-stream.ts` for deltas, timestamps, voice ids, PCM and container descriptions, and the terminal asset.

-Every modality route is the inline | stream | queued union (transcription uses all three: OpenAI and Gemini stream, Deepgram is inline, AssemblyAI is queued), every client is `MediaClient.make(Service, { modality, responseEvents })` (`src/media-client.ts`), which dispatches on the route's `kind`, and every model composes through `composeRoute`. fal queue protocols come from `protocols/utils/fal-queue.ts`, bodies are `json`, `multipart`, or `binary` (a raw upload), and a queued protocol that must upload media before submitting implements `start.prepare` (`MediaProtocol.Prepare`; AssemblyAI `/v2/upload`).
+Every modality route is the inline | stream | queued union (transcription uses all three: OpenAI and Gemini stream, Deepgram and ElevenLabs are inline, AssemblyAI is queued), every client is `MediaClient.make(Service, { modality, responseEvents })` (`src/media-client.ts`), which dispatches on the route's `kind`, and every model composes through `composeRoute`. fal queue protocols come from `protocols/utils/fal-queue.ts`, bodies are `json`, `multipart`, or `binary` (a raw upload), and a queued protocol that must upload media before submitting implements `start.prepare` (`MediaProtocol.Prepare`; AssemblyAI `/v2/upload`).

 ### URL Construction

@@ -112,7 +112,7 @@ For providers where the URL is derived from typed inputs (Azure resource name, B

 ### Provider Facades

-Provider-facing APIs are configured facades over route values. Endpoint/auth/resource/API-version setup happens before model selection, and model selectors accept only a model or deployment id. Media models use per-modality selectors on the same facade (`openai.image(id)`, later `.video` / `.speech` / `.transcription`) that mirror `openai.responses(id)`; the one-word overlap with the request namespace is accepted over a second construction path:
+Provider-facing APIs are configured facades over route values. Endpoint/auth/resource/API-version setup happens before model selection, and model selectors accept only a model or deployment id. Media models use per-modality selectors on the same facade (`openai.image(id)`, `.speech(id)`, `.transcription(id)`, `google.video(id)`) that mirror `openai.responses(id)`; the one-word overlap with the request namespace is accepted over a second construction path:

 ```ts
 const openai = OpenAI.configure({ apiKey, baseURL })
@@ -129,8 +129,9 @@ VercelAIGateway.configure().experimental.evaluation("typesafe-ai/jev")

 OpenRouter reads `OPENROUTER_API_KEY`. Vercel reads `AI_GATEWAY_API_KEY`, then `VERCEL_OIDC_TOKEN`.
 The common API uses `boolean`; System One routes lower it to native `noul`.
-Choice and score confidence plus score legends remain available in provider metadata, and the
-provider's rounded probabilities are returned unchanged.
+Choice and score answers include `confidence` when the provider returns it, such as
+`response.answers.department.confidence`. Score legends remain available in provider metadata, and
+the provider's rounded probabilities are returned unchanged.

 ## Alibaba Cloud Model Studio

@@ -475,18 +476,18 @@ const program = Effect.gen(function* () {
 Common fields are portable in shape, not in support. Unsupported fields fail with a typed `AIError` before any network
 call rather than being dropped, so check this table before swapping only the `model`:

-| Provider              | `n` | `size`    | `aspectRatio` | `seed` | `format` | `images`                  | `mask`              |
-| --------------------- | --- | --------- | ------------- | ------ | -------- | ------------------------- | ------------------- |
-| OpenAI                | ✓¹  | ✓         | ✗             | ✗      | ✓        | ✓                         | ✓                   |
-| Google (Gemini)       | 1   | ✗         | ✓             | ✓      | ✗        | ✓ (no public URLs)        | ✗                   |
-| xAI                   | ✓   | ✗         | ✓             | ✗      | ✗        | ✓                         | ✗                   |
-| Z.ai                  | ✗   | ✓         | ✗             | ✗      | ✗        | ✗                         | ✗                   |
-| Meta                  | ✓   | ✓ (hint)  | ✗             | ✗      | ✓        | ✓                         | ✗                   |
-| Black Forest Labs     | 1   | per model | per model     | ✓      | ✓        | per model (1–8)           | `flux-pro-1.0-fill` |
-| fal                   | ✓   | per model | per model     | ✓      | ✓        | 1 (several on `/edit`)    | ✓                   |
-| Replicate             | ✗   | ✗         | ✗             | ✗      | ✗        | ✗ (use `providerOptions`) | ✗                   |
-| Stability `image`     | 1   | ✗         | ✓             | ✓      | ✓        | 1 (not on `core`)         | ✗                   |
-| Stability `upscale()` | ✗   | ✗         | ✗             | ✓      | ✓        | exactly 1 (required)      | ✗                   |
+| Provider              | `n` | `size`    | `aspectRatio` | `seed` | `format` | `images`                         | `mask`              |
+| --------------------- | --- | --------- | ------------- | ------ | -------- | -------------------------------- | ------------------- |
+| OpenAI                | ✓¹  | ✓         | ✗             | ✗      | ✓        | ✓                                | ✓                   |
+| Google (Gemini)       | 1   | ✗         | ✓             | ✓      | ✗        | ✓ (no public URLs)               | ✗                   |
+| xAI                   | ✓   | ✗         | ✓             | ✗      | ✗        | ✓                                | ✗                   |
+| Z.ai                  | ✗   | ✓         | ✗             | ✗      | ✗        | ✗                                | ✗                   |
+| Meta                  | ✓   | ✓ (hint)  | ✗             | ✗      | ✓        | ✓                                | ✗                   |
+| Black Forest Labs     | 1   | per model | per model     | ✓      | ✓        | per model (1–8)                  | `flux-pro-1.0-fill` |
+| fal                   | ✓   | per model | per model     | ✓      | ✓        | 1 (several on `/edit`, `/multi`) | ✓                   |
+| Replicate             | ✗   | ✗         | ✗             | ✗      | ✗        | ✗ (use `providerOptions`)        | ✗                   |
+| Stability `image`     | 1   | ✗         | ✓             | ✓      | ✓        | 1 (not on `core`)                | ✗                   |
+| Stability `upscale()` | ✗   | ✗         | ✗             | ✓      | ✓        | exactly 1 (required)             | ✗                   |

 ✓ lowers natively; ✗ fails whenever the field is set (including `n: 1`); `1` means `n > 1` fails. ¹ `Image.stream` on OpenAI generates one image. fal
 rejects `size` and `aspectRatio` together; which one a fal or BFL model takes depends on the model.
@@ -621,8 +622,7 @@ persist the bytes promptly if they must remain available.
 ### Partial images

 OpenAI's GPT image models stream previews. `Image.stream` sends `stream: true` with `partialImages` (0–3, default 2)
-and emits `image-partial` events before each final `image`; `Image.generate` keeps the plain JSON request.
-`dall-e-*` models do not stream and fail typed:
+and emits `image-partial` events before each final `image`; `Image.generate` keeps the plain JSON request:

 ```ts
 import { Stream } from "effect"
@@ -699,7 +699,7 @@ const program = Effect.gen(function* () {
 })
 ```

-The hosted result is represented as a provider-executed tool call and tool result, and the generated image is also emitted as a first-class `media` `LLMEvent` (`response.message` then carries a `media` part). Gemini image-capable models emit the same `media` event for inline image output. Retaining `response.message` preserves the generated image for continuation on both routes.
+The hosted result is represented as a provider-executed tool call and a tool result whose content carries the generated image as a file. Gemini image-capable models instead emit a first-class `media` `LLMEvent` for inline image output (`response.message` then carries a `media` part). Retaining `response.message` preserves the generated image for continuation on both routes.

 ## Video generation

@@ -753,7 +753,10 @@ const events = Video.stream({ model: Runway.configure({ apiKey }).video("gen4.5"

 Status polls, result fetches, cancels, and asset downloads all run through the same request executor with the route's
 auth. `Generation.await` and `Generation.events` fail with a
-`Timeout` reason when `poll.timeout` (default 10 minutes) elapses. Failed,
+`Timeout` reason when `poll.timeout` (default 10 minutes) elapses. Status polls and result fetches retry transient
+failures (rate limits, provider 5xx, network errors) with backoff that honors `retry-after`, always within
+`poll.timeout`; submits and cancels never retry. Interrupting a wait (or aborting its `signal`) does not cancel the
+provider job, which keeps running and billing: call `cancel()` to stop it. Failed,
 cancelled, and expired generations fail typed with the provider's terminal document on `reason.body`; moderation
 outcomes (Veo `raiMediaFilteredReasons`, xAI `respect_moderation`, Runway `SAFETY.*` codes) surface as `notices` when
 a video is still returned and as a `ContentPolicy` reason when nothing is.
@@ -774,7 +777,9 @@ Provider notes:
 The promise client exposes the same surface: `ai.video.start(...)` resolves to a handle with `await`, `events`,
 `result`, `refresh`, `cancel`, and `token`; `ai.video.generate`, `ai.video.resume(model, token)`, and
 `ai.video.stream` mirror the Effect API. The handle's `status` and `progress` are a snapshot from when it was
-created; `refresh()` resolves to a new handle.
+created; `refresh()` resolves to a new handle. Every promise method and stream accepts `{ signal }`: like `fetch`,
+aborting rejects the Promise or throws from the `for await` loop with `signal.reason` (an `AbortError` `DOMException`
+unless `abort(reason)` passed one), while `break` stops a stream without throwing.

 ```ts
 import { ai } from "@opencode/ai/promise"
@@ -789,7 +794,7 @@ await ai.write(video.video, "./kite.mp4")

 Speech (text-to-speech) is one request whose response is parsed incrementally, so every route supports both
 `Speech.generate` (the whole file) and `Speech.stream` (audio chunks as they arrive). Models come from `.speech(...)`
-selectors on the `OpenAI`, `Google` (Gemini TTS), `ElevenLabs`, `Cartesia`, `Deepgram`, and `XAI` facades. Common fields
+selectors on the `OpenAI`, `Google` (Gemini TTS), `ElevenLabs`, `Cartesia`, and `Deepgram` facades. Common fields
 (`voice`, `format`, `speed`, `language`, `instructions`, `timestamps`) lower natively or fail with a typed `AIError`
 before any network call; provider-native controls live under `providerOptions`, inferred from the selected model.

@@ -840,9 +845,10 @@ Provider notes:
 - **OpenAI** streams over SSE (`stream_format: "sse"`), which is also the only place it reports token usage; `tts-1`
  and `tts-1-hd` do not support SSE and stream the raw audio body instead. `pcm` is 24 kHz 16-bit mono. `language`
  and `timestamps` are not supported.
- **Gemini TTS** returns raw 16-bit PCM only (`audio/L16;codec=pcm;rate=24000`), so any `format` other than `pcm`
-  fails typed; wrap the samples yourself. Style is directed in the text, so `instructions` and `speed` fail typed.
-  Only `gemini-3.1-flash-tts-preview` and later support streaming. Two-speaker audio goes through
+- **Gemini TTS** returns the provider's default output: WAV for Gemini 3.8 TTS `generate`, raw 16-bit PCM
+  (`audio/L16;codec=pcm;rate=24000`) otherwise. `pcm` is the only explicit `format` it accepts, and it fails typed on
+  Gemini 3.8 `generate`; the route never wraps PCM as WAV. Style is directed in the text, so `instructions` and
+  `speed` fail typed. Only `gemini-3.1-flash-tts-preview` and later support streaming. Two-speaker audio goes through
  `providerOptions.speechConfig.multiSpeakerVoiceConfig`.
 - **ElevenLabs** requires `voice` (the path voice id) and authenticates with `xi-api-key`. `format` maps to the
  `output_format` query parameter (`mp3_44100_128`, `pcm_24000`, `wav_24000`, `opus_48000_64`);
@@ -854,10 +860,6 @@ Provider notes:
 - **Deepgram** Aura's voice is the model id (`aura-2-thalia-en`), so `voice` and `language` fail typed. `format`
  and `providerOptions` lower to query parameters (`encoding`, `container`, `sample_rate`, `bit_rate`); `pcm` is
  `linear16` without a container. Auth is `Authorization: Token <DEEPGRAM_API_KEY>`.
- **xAI** (`POST /v1/tts`) has no model field, so the `.speech(...)` id (for example `"grok-tts"`) only names the
-  model. `voice` is the `voice_id` (default `eve`), `language` defaults to `auto`, and `format` is the codec (`mp3`,
-  `wav`, `pcm`, `mulaw`, `alaw`); `providerOptions.sampleRate` and `bitRate` complete `output_format`. Both
-  `generate` and `stream` read the raw audio body. `instructions` and `timestamps` are not supported.

 The promise client mirrors the Effect API; `ai.speech.stream` is an `AsyncIterable`.

@@ -875,11 +877,12 @@ for await (const event of ai.speech.stream({ model, text: "Hello from OpenCode."
 ## Transcription

 Transcription (speech-to-text) is the one modality whose providers use every route kind: OpenAI and Gemini stream,
-Deepgram answers inline, and AssemblyAI is queued. `Transcription.generate` and `Transcription.stream` work on all of
-them; `Transcription.start` / `resume` return a `Generation` on queued routes and fail with `UnsupportedOperation`
-elsewhere. Models come from `.transcription(...)` selectors on the `OpenAI`, `Google`, `Deepgram`, `AssemblyAI`, and
-`XAI` facades. Common fields (`language`, `prompt`, `timestamps: "none" | "segment" | "word"`, `diarize`, `speakers`) lower
-natively or fail with a typed `AIError` before any network call; a route may return more than asked.
+Deepgram and ElevenLabs answer inline, and AssemblyAI is queued. `Transcription.generate` and `Transcription.stream`
+work on all of them; `Transcription.start` / `resume` return a `Generation` on queued routes and fail with
+`UnsupportedOperation` elsewhere. Models come from `.transcription(...)` selectors on the `OpenAI`, `Google`,
+`Deepgram`, `ElevenLabs`, and `AssemblyAI` facades. Common fields (`language`, `prompt`,
+`timestamps: "none" | "segment" | "word"`, `diarize`, `speakers`) lower natively or fail with a typed `AIError` before
+any network call; a route may return more than asked.

 ```ts
 import { Console, Effect, Stream } from "effect"
@@ -891,7 +894,7 @@ const openai = OpenAI.configure({ apiKey: process.env.OPENAI_API_KEY })
 const program = Effect.gen(function* () {
  const audio = yield* Media.file("./call.mp3")

-  // Speaker-labelled segments; labels are provider-native strings ("A", "0", "spk:0").
+  // Speaker-labelled segments; labels are provider-native strings ("A", "0", "spk:0", "speaker_0").
  const response = yield* Transcription.generate({
    model: Deepgram.configure({ apiKey }).transcription("nova-3"),
    audio,
@@ -901,7 +904,7 @@ const program = Effect.gen(function* () {
  response.text // "Hello from OpenCode."
  response.segments // [{ text, startSeconds, endSeconds, speaker: "0" }]
  response.words // [{ text, startSeconds, endSeconds, speaker, confidence }]
-  response.language // the provider's own value, lowercased ("en", "english", "en_us")
+  response.language // the provider's own value, lowercased ("en", "eng", "english", "en_us")

  // Text deltas as the model transcribes, then one finish carrying the whole transcript.
  yield* Transcription.stream({ model: openai.transcription("gpt-4o-mini-transcribe"), audio }).pipe(
@@ -925,10 +928,12 @@ Provider notes:
 - **OpenAI** takes inline audio only; `diarize` needs `gpt-4o-transcribe-diarize`, timestamps need `whisper-1`, and `whisper-1` does not stream.
 - **Gemini** needs a transcribe model (`gemini-3.5-transcribe`); `prompt` and `speakers` fail typed.
 - **Deepgram** detects the language unless `language` is set; vocabulary goes in `providerOptions.keyterm`.
- **AssemblyAI** uploads inline audio before submitting and is the only route that accepts `speakers`.
- **xAI** (`grok-voice-transcribe-2.0`) answers inline and always returns words; `diarize` or `timestamps: "segment"`
-  groups them into speaker-turn segments. Vocabulary goes in `providerOptions.keyterm`, and headerless PCM uploads
-  send `audio_format` and `sample_rate` from `audio.info`. `prompt` and `speakers` fail typed.
+- **ElevenLabs** (`scribe_v2`) uploads inline audio as the multipart `file` and sends a URL as `source_url`. Words
+  always carry timestamps, and segments are speaker turns, so `diarize`, `timestamps: "segment"`, or `speakers` turns
+  on diarization. `speakers` is an upper bound (`num_speakers`); `prompt` fails typed (vocabulary goes in
+  `providerOptions.keyterms`), as do webhook delivery and per-channel output (`use_multi_channel` without
+  `multichannel_output_style: "combined"`).
+- **AssemblyAI** uploads inline audio before submitting and treats `speakers` as the exact speaker count.

 The promise client mirrors the Effect API:

@@ -951,6 +956,7 @@ const transcript = await generation.await({ poll: { interval: 3_000 } })
 - **`ImageClient`** — Effect service and layer for image execution, parallel to `LLMClient`.
 - **`Media`** — the shared asset type (`Media.Asset`, `Media.Source`) and constructors used by messages, tool results, and media requests.
 - **`Generation`** — provider-neutral handle for an in-flight media generation (`await`, `refresh`, `cancel`, `events`) used by queued media routes.
+- **`Video.request` / `generate` / `stream` / `start` / `resume`** — queued video generation through a provider-neutral request; `VideoClient` is its Effect service and layer.
 - **`Speech.request` / `Speech.generate` / `Speech.stream`** — text-to-speech through a provider-neutral request; `SpeechClient` is its Effect service and layer.
 - **`Transcription.request` / `generate` / `stream` / `start` / `resume`** — speech-to-text over inline, streaming, and queued routes; `TranscriptionClient` is its Effect service and layer.
 - **`AIClient.layer` / `AIClient.layerWith(executor)`** — every modality client plus the request executor in one layer.
@@ -1217,7 +1223,7 @@ const gateway = CloudflareAIGateway.configure({
 }).model("workers-ai/@cf/meta/llama-3.1-8b-instruct")
 ```

-Included LLM providers: OpenAI, Anthropic, Google (Gemini), Google Vertex, Amazon Bedrock, Azure OpenAI, Baseten, Cerebras, Cloudflare AI Gateway, Cloudflare Workers AI, DeepInfra, DeepSeek, Fireworks, Groq, Mistral, OpenRouter, TogetherAI, and xAI. Z.ai currently exposes image generation. Generic Chat Completions, Responses, and Anthropic Messages-compatible entrypoints support custom endpoints.
+Included LLM providers: OpenAI, Anthropic, Google (Gemini), Google Vertex, Amazon Bedrock, Azure OpenAI, Baseten, Cerebras, Cohere, Cloudflare AI Gateway, Cloudflare Workers AI, DeepInfra, DeepSeek, Fireworks, Groq, Mistral, OpenRouter, TogetherAI, and xAI. Z.ai currently exposes image generation. Generic Chat Completions, Responses, and Anthropic Messages-compatible entrypoints support custom endpoints.

 Each named provider owns its module, endpoint, authentication, and route setup. Providers with the same wire format compose the shared protocol directly:

@@ -40,7 +40,7 @@ The design below is derived from a survey of the raw provider APIs (OpenAI, Gemi

 ### Model selection

-A model value is built as `OpenAI.configure({ apiKey }).responses("gpt-5")` or `.image("gpt-image-2")`: `configure` fixes credentials, endpoint, and defaults; the selector fixes which of the provider's APIs to hit and binds the typed `providerOptions` generic. Media follows the same shape with one selector per modality — `openai.image(id)` today, `.video(id)` / `.speech(id)` / `.transcription(id)` as those modalities land — mirroring `openai.responses(id)`. `Image.request` accepts `ImageModel` only, exactly as `LLM.request` accepts `LanguageModel`.
+A model value is built as `OpenAI.configure({ apiKey }).responses("gpt-5")` or `.image("gpt-image-2")`: `configure` fixes credentials, endpoint, and defaults; the selector fixes which of the provider's APIs to hit and binds the typed `providerOptions` generic. Media follows the same shape with one selector per modality — `.image(id)`, `.video(id)`, `.speech(id)`, `.transcription(id)` on the facades that offer each — mirroring `openai.responses(id)`. `Image.request` accepts `ImageModel` only, exactly as `LLM.request` accepts `LanguageModel`.

 ```ts
 import { OpenAI, Google } from "@opencode/ai/providers"
@@ -135,8 +135,8 @@ Effect.gen(function* () {
 })
 ```

-`size` and `aspectRatio` are not interchangeable; each route rejects fields it cannot lower — see the README's Image
-portability matrix.
+`size` and `aspectRatio` are not interchangeable; each route rejects fields it cannot lower — see the portability table
+in the README's Image generation section.

 Editing is not a separate function; `images`/`mask` on the request select the edit path in the route (OpenAI `/images/edits`, Gemini multimodal parts, xAI `/images/edits`). Routes that cannot honor `mask` fail with `Unsupported`.

@@ -166,8 +166,8 @@ Effect.gen(function* () {

  // Simple: wait for it.
  const response = yield* Video.generate(request, { poll: { interval: "10 seconds", timeout: "10 minutes" } })
-  response.video                                     // Media.Asset: url with expiresAt (+ transient `headers` for Veo downloads)
-  response.usage                                     // credits on Runway; the other three report none
+  response.video                                     // Media.Asset: url (expiresAt on Veo and Runway; transient `headers` for Veo downloads)
+  response.usage                                     // credits on Runway; the other three report none (xAI's usage.cost_in_usd_ticks is not decoded)
  response.notices                                   // Veo raiMediaFilteredReasons → filtered, xAI respect_moderation → moderated
  yield* response.video.materialize()                // pull bytes before the URL expires

@@ -175,7 +175,7 @@ Effect.gen(function* () {
  const generation = yield* Video.start(request)     // Generation<VideoResponse>
  generation.id; generation.status; generation.progress; generation.position; generation.token
  yield* generation.await({ poll })                  // VideoResponse
-  yield* generation.cancel()                         // fal PUT cancel_url, Runway DELETE /tasks/{id}; no-op for Veo and xAI
+  yield* generation.cancel()                         // fal PUT cancel_url, Runway DELETE /tasks/{id}; Veo and xAI succeed without a request

  // Resume from another process. The token is validated against the route's codec and refreshed once. It carries no
  // route identity, so persist the provider and model ID alongside it: `resume` needs the model.
@@ -188,10 +188,11 @@ Effect.gen(function* () {

 Tokens are route-owned JSON: Veo `{ operation }`, xAI `{ requestID }`, Runway `{ taskID }`, fal
 `{ requestID, statusURL, responseURL, cancelURL }` (fal's follow-up URLs are authoritative and absolute). Common-field
-lowering per provider: Veo takes inline media only and rejects `audio: false` and `n > 1`; xAI rejects `seed` and
-`negativePrompt` and routes a `video` input to edits or (`providerOptions.mode: "extend"`) extensions; fal rejects
-`durationSeconds`, `references`, and `frames.last` because the field names and enums differ per model; Runway passes
-`aspectRatio` through as its pixel `ratio` and rejects `n`.
+lowering per provider: Veo takes inline media only, rejects `audio: false` and `n > 1`, and requires `frames.first`
+when `frames.last` is set; xAI rejects `n`, `seed`, and `negativePrompt` and routes a `video` input to edits or
+(`providerOptions.mode: "extend"`) extensions; fal rejects `n`, plus `durationSeconds`, `references`, and `frames.last`
+because the field names and enums differ per model; Runway passes `aspectRatio` through as its pixel `ratio` and
+rejects `n`.

 Deferred: `Video.complete(model, token, webhook)` (finish from a webhook payload without polling) and provider poll
 hints (none of the four providers emit one). Later providers: Luma, Kling, MiniMax, Replicate.
@@ -199,8 +200,7 @@ hints (none of the four providers emit one). Later providers: Luma, Kling, MiniM
 #### Speech (TTS)

 Shipped in phase 3 (`src/speech.ts`, `src/speech-client.ts`, protocols `openai-speech`, `google-speech`,
-`elevenlabs-speech`, `cartesia-speech`, `deepgram-speech`, `xai-speech`; new `ElevenLabs`, `Cartesia`, and `Deepgram`
-facades).
+`elevenlabs-speech`, `cartesia-speech`, `deepgram-speech`; new `ElevenLabs`, `Cartesia`, and `Deepgram` facades).

 ```ts
 const request = Speech.request({
@@ -242,14 +242,16 @@ name→id resolution. Multi-speaker (Gemini `speechConfig.multiSpeakerVoiceConfi
 `opus_48000_64`, Cartesia `{ container, encoding, sample_rate }`, Deepgram `encoding`+`container`) and declares the
 asset's media type rather than sniffing, because headerless PCM can look like an MPEG frame sync. Headerless PCM
 always carries `info.encoding`, `info.sampleRate`, and `info.channels`; its media type is the provider's declaration
-(Gemini `audio/L16;codec=pcm;rate=24000`, Deepgram's `content-type`) or `audio/pcm`. Gemini returns PCM only, so any
-other `format` is rejected rather than wrapped as WAV by the route. Every `format` value a route cannot produce (unknown
-to it, a container on Cartesia SSE, WAV on an ElevenLabs stream, anything but PCM on Gemini) fails the same way as an
-unsupported field: `UnsupportedOperation` with `operation: "media.format"`.
+(Gemini `audio/L16;codec=pcm;rate=24000`, Deepgram's `content-type`) or `audio/pcm`. Gemini's asset follows the
+provider's declared type: WAV for Gemini 3.8 TTS `generate`, headerless PCM otherwise. The route never wraps PCM as WAV,
+so `pcm` is the only explicit `format` it accepts, and not on Gemini 3.8 `generate`. Every `format` value a route cannot
+produce (unknown to it, a container on Cartesia SSE, WAV on an ElevenLabs stream, anything but `pcm` on Gemini, `pcm` on
+Gemini 3.8 `generate`) fails the same way as an unsupported field: `UnsupportedOperation` with
+`operation: "media.format"`.

 **Timestamps.** `timestamps: true` on the request asks for alignment. ElevenLabs selects the `with-timestamps`
 endpoints (character-level, NDJSON when streaming); Cartesia sets `add_timestamps` on `/tts/sse` (word-level; a
-`generate` with timestamps collects the SSE stream). OpenAI, Gemini, Deepgram, and xAI reject it.
+`generate` with timestamps collects the SSE stream). OpenAI, Gemini, and Deepgram reject it.

 Common-field lowering per provider:

@@ -260,7 +262,6 @@ Common-field lowering per provider:
 | ElevenLabs | path voice id (required) | `voice_settings.speed` | `language_code` | unsupported | `with-timestamps` | `credits` from `character-cost` header |
 | Cartesia | `voice` (required) | `generation_config.speed` | `language` | unsupported | `add_timestamps` | none |
 | Deepgram | unsupported (voice is the model) | `speed` query | unsupported | unsupported | unsupported | `characters` from `dg-char-count` header |
-| xAI | `voice_id` (defaults to `eve`) | `speed` | `language` (`auto` when omitted) | unsupported | unsupported | none |

 Deferred: `Speech.session(...)` — input-streaming TTS where text arrives incrementally over a WebSocket (ElevenLabs
 `stream-input`, Cartesia WebSocket contexts, Deepgram WebSocket speak) — is a separate scoped resource, not part of
@@ -269,8 +270,8 @@ Deferred: `Speech.session(...)` — input-streaming TTS where text arrives incre
 #### Transcription (STT)

 Shipped as the second half of phase 3 (`src/transcription.ts`, `src/transcription-client.ts`, protocols
-`openai-transcription`, `google-transcription`, `deepgram-transcription`, `assemblyai-transcription`,
-`xai-transcription`; new `AssemblyAI` facade).
+`openai-transcription`, `google-transcription`, `deepgram-transcription`, `elevenlabs-transcription`,
+`assemblyai-transcription`; new `AssemblyAI` facade).

 ```ts
 const request = Transcription.request({
@@ -279,7 +280,7 @@ const request = Transcription.request({
  language: "en",                                  // provider-native passthrough
  timestamps: "segment",                           // none | segment | word
  diarize: true,
-  speakers: 2,                                     // expected count, hint only (AssemblyAI)
+  speakers: 2,                                     // speaker count (AssemblyAI exact, ElevenLabs maximum)
  providerOptions: { known_speaker_names: ["agent"] },
 })

@@ -293,10 +294,11 @@ yield* Transcription.resume(model, token)
 Transcription is the first modality whose providers span all three protocol kinds, and it needed no fourth kind.
 Every `MediaRoute` now carries its `kind`; `TranscriptionRoute` is the union of the inline, stream, and queued routes;
 `TranscriptionModel.fromRoute` is overloaded per protocol kind (arity picks the overload: `<Options>`,
-`<Options, Frame, State>`, `<Options, Token>`) and composes through `MediaRoute.inline` / `stream` / `queued`; and
-`TranscriptionClient` dispatches on `route.kind`. `generate` on a queued route is `start` then `await`; `stream` on an
-inline route is the response as a single `finish`, and on a queued route it is the status observations followed by
-`finish`. `start` / `resume` on a non-queued route fail with `UnsupportedOperation` (`transcription.start`). The
+`<Options, Frame, State>`, `<Options, Token>`) and composes through the shared `composeRoute` (`src/media-model.ts`),
+which picks `MediaRoute.inline` / `stream` / `queued`; and `TranscriptionClient`, like every modality client, is
+`MediaClient.make` (`src/media-client.ts`), which dispatches on `route.kind`. `generate` on a queued route is `start`
+then `await`; `stream` on an inline route is the response as a single `finish`, and on a queued route it is the status
+observations followed by `finish`. `start` / `resume` on a non-queued route fail with `UnsupportedOperation` (`transcription.start`). The
 `finish` event carries the whole transcript (text, segments, words, language, duration, usage), so the stream route's
 `collect` is just "take `finish`".

@@ -306,16 +308,23 @@ upload); `packages/ai/AGENTS.md` (Media Routes) describes both.
 Settled rules:

 - **Timestamps.** A granularity the selected route or model cannot produce fails as `UnsupportedOperation`
-  (`media.timestamps`), following Speech; a route that returns more than asked (Deepgram and AssemblyAI always return
-  words) is not stripped. Segments always carry start and end times: Gemini times each transcription part from its
+  (`media.timestamps`), following Speech; a route that returns more than asked (Deepgram, ElevenLabs, and AssemblyAI
+  always return words) is not stripped. Segments always carry start and end times: Gemini times each transcription part from its
  word offsets, so segment timestamps and diarization also request word offsets there.
 - **Diarization.** `diarize` means segments (and words, where the provider labels them) carry `speaker`. Labels are
-  provider-native strings — OpenAI `A` or a known speaker name, Deepgram `0`, Gemini `spk:0`, AssemblyAI `A` — with no
-  cross-provider speaker model. `speakers` is a hint; only AssemblyAI (`speakers_expected`) accepts it.
+  provider-native strings — OpenAI `A` or a known speaker name, Deepgram `0`, Gemini `spk:0`, AssemblyAI `A`,
+  ElevenLabs `speaker_0` — with no cross-provider speaker model. `speakers` is the number of speakers to label:
+  AssemblyAI (`speakers_expected`) treats it as an exact constraint rather than a hint, and ElevenLabs
+  (`num_speakers`) as the maximum. Both turn on diarization for it; the other routes reject it.
+- **Segments from words.** ElevenLabs returns only a token list (`word`, `spacing`, `audio_event`), so its segments
+  are speaker turns: consecutive words and spacing with one `speaker_id`, text joined from the provider's own spacing
+  tokens. `words` drops spacing and audio events. Segments therefore need diarization, which `timestamps: "segment"`
+  turns on, as AssemblyAI's utterances need speaker labels.
 - **Language** is passed through (`language`, OpenAI `gpt-transcribe` `languages[]`, Gemini `languageCodes`,
-  AssemblyAI `language_code`). `response.language` is the provider's own value, lowercased but not normalized: an
-  ISO code on most routes, `english` from whisper-1, `en_us` from AssemblyAI. Deepgram and AssemblyAI assume English
-  unless asked to detect, so a missing `language` enables their detection.
+  AssemblyAI and ElevenLabs `language_code`). `response.language` is the provider's own value, lowercased but not
+  normalized: an ISO code on most routes (AssemblyAI's detection returns `en`, ElevenLabs ISO 639-3 `eng`), `english`
+  from whisper-1. Deepgram and AssemblyAI assume English unless asked to detect, so a missing `language` enables their
+  detection.
 - **Gemini** requires a transcribe model; other model ids fail with `UnsupportedOperation` before the call, because
  general models ignore `audioTranscriptionConfig` and answer conversationally. Streamed chunks carry whole speaker
  turns (one part per turn), which join with a space.
@@ -325,15 +334,15 @@ Settled rules:

 | Provider | Kind | Audio input | `timestamps` | `diarize` | Unsupported | Usage |
 |---|---|---|---|---|---|---|
-| OpenAI | stream (`stream: true` in `stream` mode) | multipart `file` (inline only) | `whisper-1` (`verbose_json`); diarize model: `segment` | `gpt-4o-transcribe-diarize` (`diarized_json`) | `speakers`; `prompt` on the diarize model; streaming on `whisper-1` | `tokens` or `seconds` |
+| OpenAI | stream (`stream: true` in `stream` mode; `whisper-1` ignores `stream`, so it emits only `finish`) | multipart `file` (inline only) | `whisper-1` (`verbose_json`); diarize model: `segment` | `gpt-4o-transcribe-diarize` (`diarized_json`) | `speakers`; `prompt` on the diarize model | `tokens` or `seconds` |
 | Gemini | stream (`generateContent` / `streamGenerateContent`) | `inlineData` or Gemini Files `fileData` | `audioTranscriptionConfig.wordTimestamp` | `audioTranscriptionConfig.diarization` | `prompt`, `speakers` | `tokens` |
 | Deepgram | inline | raw body, or JSON `{ url }` | words always; `segment` → `utterances` | `diarize_model=latest` + `utterances` | `prompt`, `speakers` | `seconds` (`metadata.duration`) |
+| ElevenLabs | inline | multipart `file`, or `source_url` | words always; `segment` → `diarize` (speaker turns) | `diarize` | `prompt`; `webhook`, per-channel `use_multi_channel` | `seconds` (`audio_duration_secs`) |
 | AssemblyAI | queued (upload → submit → poll) | `/v2/upload` then `audio_url`, or a URL | words always; `segment` → `speaker_labels` | `speaker_labels` | — | `seconds` (`audio_duration`) |
-| xAI | inline (batch `/v1/stt`) | multipart `file` (last field), or `url` | words always; `segment` → `diarize` speaker turns | `diarize` | `prompt`, `speakers` | `seconds` (`duration`) |

 Deferred: `Transcription.session(...)` — realtime STT over WebSocket (Deepgram live, AssemblyAI streaming, ElevenLabs
 realtime, OpenAI realtime transcription) — is the same future scoped `session` shape as input-streaming TTS and ships
-with the realtime work in phase 5. ElevenLabs Scribe is not implemented yet.
+with the realtime work in phase 5.

 ### `Generation` — shared async execution

@@ -356,7 +365,11 @@ GenerationAwaitOptions = { poll?: Poll }
 Poll = { interval?: Duration; timeout?: Duration }
 ```

-`Generation` is not video-specific. Image routes on BFL, fal, and Replicate are queued; `Image.start` exists for them. A route declares itself `inline` or `queued`; `generate` on a queued route is `start` then `await`.
+`Generation` is not video-specific. Image routes on BFL, fal, Replicate, and Stability `upscale()` are queued; `Image.start` exists for them. A route declares itself `inline` or `queued`; `generate` on a queued route is `start` then `await`.
+
+Status polls and result reads retry transient failures (rate limits, provider 5xx, and transport errors, classified by the same `isRetryable` the Session runner uses) inside `MediaRoute.queued`. Only the HTTP exchange retries, never the decoded document: a terminal `failed` generation also surfaces as `ProviderInternal` and must not be re-read. Gaps grow exponentially from 1s with jitter, up to 30s each, honoring a provider `retry-after` up to that cap, for at most 8 retries. `await`, `events`, and `Video.stream` cut retries off at `poll.timeout` and fail with `Timeout`, so retries never extend the caller's deadline; a direct `result()` or `resume` read is bounded by the retry cap alone. `start` and `cancel` never retry: a repeated submit can start and bill a second job. The policy is internal; there is no option for it.
+
+Interrupting `await`, `events`, or `Video.stream` (or aborting the promise API's `signal`) stops waiting only. The provider job keeps running and billing; call `cancel()` explicitly to stop it.

 ### Usage

@@ -399,18 +412,19 @@ for await (const event of ai.llm.stream(request)) { … }
 await ai.dispose()
 ```

-Streams become `AsyncIterable` via `Stream.toAsyncIterable`. `AIError` is thrown as-is. `AbortSignal` maps to interruption. Nothing in `src/*` except this entrypoint knows about promises.
+Streams become `AsyncIterable` via `Stream.toAsyncIterable`. `AIError` is thrown as-is. Aborting an `AbortSignal` interrupts the work and, like `fetch`, rejects the Promise or throws from the stream with `signal.reason` instead of ending the stream as if complete. Nothing in `src/*` except this entrypoint knows about promises.

 ### Providers

-Existing facades gain per-modality selectors; the modality routes each facade provides:
+Existing facades gain per-modality selectors; the modality routes each facade provides (*italics* are not
+implemented):

 | Facade | llm | image | video | speech | transcription | other |
 |---|---|---|---|---|---|---|
-| `OpenAI` | responses (default), chat | Images API (stream) | Sora (deprecated 2026-09-24) | ✓ | ✓ | |
+| `OpenAI` | responses (default), chat | Images API (stream) | *Sora skipped (decision 8)* | ✓ | ✓ | |
 | `Google` | Gemini | Gemini-native | Veo | Gemini TTS | `gemini-3.5-transcribe` | |
-| `XAI` | ✓ | ✓ | ✓ | ✓ | ✓ (batch) | |
-| `ElevenLabs` | | | | ✓ | Scribe | soundEffect, music |
+| `XAI` | ✓ | ✓ | ✓ | | | |
+| `ElevenLabs` | | | | ✓ | Scribe | *soundEffect, music (phase 5)* |
 | `Cartesia` | | | | ✓ | | |
 | `Deepgram` | | | | Aura | ✓ | |
 | `Fal` | | ✓ (queued) | ✓ | | | |
@@ -419,7 +433,7 @@ Existing facades gain per-modality selectors; the modality routes each facade pr
 | `Replicate` | | ✓ (queued) | | | | |
 | `Stability` | | `image` (inline), `upscale()` (queued) | | | | |
 | `Runway` | | | ✓ | | | |
-| `Luma`, `Kling`, `MiniMax` | | per provider | | | | |
+| `Luma`, `Kling`, `MiniMax` | | *deferred* | *deferred* | | | |

 New facades follow the existing one-file-per-provider rule. The facade selector is the public path for media models; modality-specific package entrypoints (for example `@opencode/ai/providers/openai/images`) are deferred until Core has a modality-aware model resolver.

@@ -463,7 +477,7 @@ Foundation + Image ship together as the reference implementation, serially. Vide

 1. **Foundation** — per-modality selectors, `Media`, `Generation`, `Poll`, `Usage` union, `MediaProtocol` kinds, `@opencode/ai/promise` with `llm` + `image`. Port the five existing image protocols onto it. Unify `MediaPart` and add the `media` LLM event (fixes Gemini image output being dropped).
 2. **Video** — ✅ Veo, xAI, fal, Runway shipped (`MediaProtocol.queued`, `Video.start/generate/resume/stream`, promise `ai.video`). Deferred: `Video.complete` (webhooks), Luma, Kling, MiniMax, Replicate.
-3. **Speech + Transcription** — ✅ Speech: OpenAI, Gemini TTS, ElevenLabs, Cartesia, Deepgram, xAI shipped (`MediaProtocol.stream`, `Speech.generate/stream`, promise `ai.speech`). ✅ Transcription: OpenAI, Gemini, Deepgram, AssemblyAI, xAI shipped across all three route kinds (`Transcription.generate/stream/start/resume`, promise `ai.transcription`). Pending: ElevenLabs Scribe. Deferred: `Speech.session` and `Transcription.session` (WebSocket streaming).
+3. **Speech + Transcription** — ✅ Speech: OpenAI, Gemini TTS, ElevenLabs, Cartesia, Deepgram shipped (`MediaProtocol.stream`, `Speech.generate/stream`, promise `ai.speech`). ✅ Transcription: OpenAI, Gemini, Deepgram, ElevenLabs Scribe, AssemblyAI shipped across all three route kinds (`Transcription.generate/stream/start/resume`, promise `ai.transcription`). Deferred: `Speech.session` and `Transcription.session` (WebSocket streaming).
 4. **Image queued routes and partials** — ✅ BFL, fal, Replicate, and Stability creative upscale queued; Stability generate inline; OpenAI `partial_images` streaming (`image-partial` restored). Imagen dropped: shut down on the Gemini API and discontinued on Vertex (2026-06-30). Deferred: Stability's synchronous edit and fast/conservative upscale endpoints.
 5. **Later** — ElevenLabs music/SFX, Lyria, `Speech.session` / `Transcription.session`, realtime.

@@ -1,6 +1,6 @@
 {
  "$schema": "https://json.schemastore.org/package.json",
-  "version": "2.0.16",
+  "version": "2.0.22",
  "name": "@opencode/ai",
  "type": "module",
  "license": "MIT",
@@ -38,12 +38,33 @@ const resolve = (policy: CachePolicy | undefined): CachePolicyObject => {
 // prefix caching, Gemini's implicit + out-of-band CachedContent). Skip the
 // whole policy pass for these — emitting hints would be harmless but pointless.
 const RESPECTS_INLINE_HINTS = new Set([
+  "alibaba-chat",
+  "alibaba-messages",
  "anthropic-messages",
+  "anthropic-compatible-messages",
+  "cloudflare-ai-gateway-messages",
  "google-vertex-messages",
+  "meta-messages",
+  "minimax-messages",
+  "moonshot-messages",
+  "zai-coding-messages",
  "bedrock-converse",
  "openrouter",
+  "digitalocean",
 ])

+// OpenRouter upstreams other than Anthropic and Alibaba Qwen cache without breakpoints. Gemini uses only the last
+// breakpoint, so a conversation-tail breakpoint writes a new cache every step and costs more than none. Qwen ignores
+// breakpoints on tool definitions and caches tools with the system prompt.
+const QWEN: CachePolicyObject = { system: true, messages: { tail: 1 } }
+const openRouterPolicy = (modelID: string): CachePolicyObject => {
+  // `~anthropic/claude-sonnet-latest` style IDs are OpenRouter aliases for the latest model in a family.
+  const id = modelID.replace(/^~/, "")
+  if (id.startsWith("anthropic/")) return AUTO
+  if (id.startsWith("qwen/")) return QWEN
+  return NONE
+}
+
 const makeHint = (ttlSeconds: number | undefined): CacheHint =>
  ttlSeconds !== undefined ? new CacheHint({ type: "ephemeral", ttlSeconds }) : new CacheHint({ type: "ephemeral" })

@@ -149,9 +170,14 @@ const countHints = (request: LLMRequest) =>

 export const applyCachePolicy = (request: LLMRequest): LLMRequest => {
  if (!RESPECTS_INLINE_HINTS.has(request.model.route.id)) return request
-  if (request.model.route.id === "openrouter" && (request.cache === undefined || request.cache === "auto"))
-    return request
-  const policy = resolve(request.cache)
+  const policy =
+    request.model.route.id === "openrouter" && (request.cache === undefined || request.cache === "auto")
+      ? openRouterPolicy(request.model.id)
+      : request.model.route.id === "alibaba-chat" && (request.cache === undefined || request.cache === "auto")
+        ? request.model.id.toLowerCase().startsWith("qwen")
+          ? QWEN
+          : NONE
+        : resolve(request.cache)
  if (!policy.tools && !policy.system && !policy.messages) return request

  const hint = makeHint(policy.ttlSeconds)
@@ -63,6 +63,7 @@ export const ChoiceAnswer = Schema.Struct({
  type: Schema.Literal("choice"),
  choice: Schema.String,
  probabilities: Schema.optional(Schema.Record(Schema.String, Probability)),
+  confidence: Schema.optional(Probability),
 })
 export type ChoiceAnswer = Schema.Schema.Type<typeof ChoiceAnswer>

@@ -70,6 +71,7 @@ export const ScoreAnswer = Schema.Struct({
  type: Schema.Literal("score"),
  score: Schema.Number,
  probabilities: Schema.optional(Schema.Record(Schema.String, Probability)),
+  confidence: Schema.optional(Probability),
 })
 export type ScoreAnswer = Schema.Schema.Type<typeof ScoreAnswer>

@@ -92,6 +94,7 @@ export type AnswerFor<Question extends EvaluationQuestion> = Question extends {
      readonly type: "choice"
      readonly choice: Extract<keyof Criteria, string>
      readonly probabilities?: Readonly<Record<Extract<keyof Criteria, string>, number>>
+      readonly confidence?: number
    }
  : Question extends { readonly type: "score" }
    ? ScoreAnswer
@@ -142,32 +142,37 @@ export const model = <Options extends EvaluationOptions = EvaluationOptions>(cfg
            Effect.mapError((cause) => fail("System One returned an invalid response", cause, text)),
          )

-          const confidence: Record<string, number> = {}
          const legend: Record<string, Record<string, Schema.Json>> = {}
          const answers = Object.fromEntries(
            Object.entries(data.answers).map(([id, answer]): [string, EvaluationAnswer] => {
              if (answer.type === "noul") return [id, { type: "boolean", probability: answer.noul }]
              if (answer.type === "choice") {
-                if (answer.confidence !== undefined) confidence[id] = answer.confidence
                return [
                  id,
                  {
                    type: "choice",
                    choice: answer.choice,
                    probabilities: answer.probabilities,
+                    ...(answer.confidence === undefined ? {} : { confidence: answer.confidence }),
                  },
                ]
              }
-              if (answer.confidence !== undefined) confidence[id] = answer.confidence
              if (answer.legend !== undefined) legend[id] = answer.legend
-              return [id, { type: "score", score: answer.score, probabilities: answer.probabilities }]
+              return [
+                id,
+                {
+                  type: "score",
+                  score: answer.score,
+                  probabilities: answer.probabilities,
+                  ...(answer.confidence === undefined ? {} : { confidence: answer.confidence }),
+                },
+              ]
            }),
          )
          const meta = {
            ...(data.id === undefined ? {} : { responseId: data.id }),
            ...(data.provider === undefined ? {} : { provider: data.provider }),
            ...data.provider_metadata?.[cfg.providerMetadataKey],
-            ...(Object.keys(confidence).length === 0 ? {} : { confidence }),
            ...(Object.keys(legend).length === 0 ? {} : { legend }),
          }
          return new EvaluationResponse({
@@ -53,6 +53,8 @@ export type Event = Observation | { readonly type: "generation-finished"; readon

 const TERMINAL: ReadonlySet<Status> = new Set(["completed", "failed", "cancelled", "expired"])

+export const isTerminal = (status: Status) => TERMINAL.has(status)
+
 export class Generation<Response> {
  readonly id: string
  readonly status: Status
@@ -81,7 +83,7 @@ export class Generation<Response> {
  }

  get terminal() {
-    return TERMINAL.has(this.status)
+    return isTerminal(this.status)
  }

  refresh(): Effect.Effect<Generation<Response>, AIError> {
@@ -100,7 +102,7 @@ export class Generation<Response> {
    return settled.pipe(
      // Non-completed terminal states also go through `result` so the route can surface its provider failure body.
      Effect.flatMap((generation) => generation.result()),
-      Effect.timeoutOrElse({ duration: timeout, orElse: () => this.timeoutError(timeout) }),
+      Effect.timeoutOrElse({ duration: timeout, orElse: () => timeoutError(this.id, timeout) }),
    )
  }

@@ -109,9 +111,10 @@ export class Generation<Response> {
  }

  /**
-   * Status observations as a stream, ending after the first terminal observation. Each poll is bounded by the time
-   * remaining until `poll.timeout`, so a hung status request fails the stream instead of stalling it. (`Stream.interruptWhen`
-   * would express this directly but deadlocks under `TestClock` when the source completes while the timer sleeps.)
+   * Status observations as a stream, ending after the first terminal observation. Each poll and each sleep between polls
+   * is bounded by the time remaining until `poll.timeout`, so a hung status request or a long interval fails the stream at
+   * the deadline instead of stalling it. (`Stream.interruptWhen` would express this directly but deadlocks under
+   * `TestClock` when the source completes while the timer sleeps.)
   */
  events(options?: AwaitOptions): Stream.Stream<Event, AIError> {
    if (this.terminal) return Stream.make(this.event())
@@ -120,17 +123,13 @@ export class Generation<Response> {
      Clock.currentTimeMillis.pipe(
        Effect.map((start) => {
          const deadline = start + Duration.toMillis(timeout)
-          const refresh = Clock.currentTimeMillis.pipe(
-            Effect.flatMap((now) =>
-              this.refresh().pipe(
-                Effect.timeoutOrElse({
-                  duration: Duration.millis(Math.max(0, deadline - now)),
-                  orElse: () => this.timeoutError(timeout),
-                }),
-              ),
+          const refresh = within(this.refresh(), this.id, timeout, deadline)
+          const schedule = this.schedule(options?.poll).pipe(
+            Schedule.modifyDelay((meta) =>
+              Effect.succeed(Duration.min(meta.duration, Duration.millis(Math.max(0, deadline - meta.now)))),
            ),
          )
-          return Stream.fromEffectSchedule(refresh, this.schedule(options?.poll)).pipe(
+          return Stream.fromEffectSchedule(refresh, schedule).pipe(
            Stream.takeUntil((generation) => generation.terminal),
            Stream.map((generation) => generation.event()),
          )
@@ -145,15 +144,6 @@ export class Generation<Response> {
    return { type: "generation-progress", id: this.id, progress: this.progress }
  }

-  private timeoutError(timeout: Duration.Duration) {
-    return new AIError({
-      reason: new TimeoutError({
-        message: `Generation ${this.id} did not finish within ${Duration.format(timeout)}`,
-        timeoutMs: Duration.toMillis(timeout),
-      }),
-    })
-  }
-
  private poll(poll: Poll | undefined) {
    return this.refresh().pipe(
      Effect.repeat({ schedule: this.schedule(poll), until: (generation) => generation.terminal }),
@@ -165,12 +155,53 @@ export class Generation<Response> {
  }
 }

+/** `events` followed by the expanded result, with the result fetch bounded by the same `poll.timeout` deadline. */
 export const resultEvents = <Response, A>(
  generation: Generation<Response>,
  expand: (response: Response) => ReadonlyArray<A>,
  options?: AwaitOptions,
-): Stream.Stream<Observation | A, AIError> =>
-  generation.events(options).pipe(
-    Stream.filter((event): event is Observation => event.type !== "generation-finished"),
-    Stream.concat(Stream.fromIterableEffect(Effect.map(generation.result(), expand))),
+): Stream.Stream<Observation | A, AIError> => {
+  const timeout = Duration.fromInputUnsafe(options?.poll?.timeout ?? DEFAULT_POLL_TIMEOUT)
+  return Stream.unwrap(
+    Clock.currentTimeMillis.pipe(
+      Effect.map((start) =>
+        generation.events(options).pipe(
+          Stream.filter((event): event is Observation => event.type !== "generation-finished"),
+          Stream.concat(
+            Stream.fromIterableEffect(
+              within(generation.result(), generation.id, timeout, start + Duration.toMillis(timeout)).pipe(
+                Effect.map(expand),
+              ),
+            ),
+          ),
+        ),
+      ),
+    ),
  )
+}
+
+/**
+ * Run `effect` within the time left until `deadline`. Fails before starting once the deadline has passed: a fast
+ * request could otherwise win the zero-budget race and schedule another zero-delay poll.
+ */
+const within = <A>(effect: Effect.Effect<A, AIError>, id: string, timeout: Duration.Duration, deadline: number) =>
+  Clock.currentTimeMillis.pipe(
+    Effect.flatMap((now) =>
+      now >= deadline
+        ? Effect.fail(timeoutError(id, timeout))
+        : effect.pipe(
+            Effect.timeoutOrElse({
+              duration: Duration.millis(deadline - now),
+              orElse: () => Effect.fail(timeoutError(id, timeout)),
+            }),
+          ),
+    ),
+  )
+
+const timeoutError = (id: string, timeout: Duration.Duration) =>
+  new AIError({
+    reason: new TimeoutError({
+      message: `Generation ${id} did not finish within ${Duration.format(timeout)}`,
+      timeoutMs: Duration.toMillis(timeout),
+    }),
+  })
@@ -4,7 +4,7 @@ export { ImageClient } from "./image-client.js"
 export { Auth } from "./route/auth.js"
 export { Provider } from "./provider.js"
 export { ProviderPackage } from "./provider-package.js"
-export { isContextOverflow, isContextOverflowFailure } from "./provider-error.js"
+export { isContextOverflow, isContextOverflowFailure, isRetryable } from "./provider-error.js"
 export type {
  RouteLanguageModelInput,
  RouteRoutedLanguageModelInput,
@@ -42,7 +42,7 @@ export type GenerationHandle<Response> = Snapshot & {
  /** Serializable JSON; pass it back to `resume` from another process. */
  readonly token: unknown
  readonly await: (options?: AwaitOptions & RunOptions) => Promise<Response>
-  /** Status observations until the first terminal one, polling like `await`; abort ends iteration without throwing. */
+  /** Status observations until the first terminal one, polling like `await`; abort throws `signal.reason`. */
  readonly events: (options?: AwaitOptions & RunOptions) => AsyncIterable<Event>
  /** The result without polling; fails when the generation has not completed. */
  readonly result: (options?: RunOptions) => Promise<Response>
@@ -50,15 +50,16 @@ export type GenerationHandle<Response> = Snapshot & {
  readonly cancel: (options?: RunOptions) => Promise<void>
 }

+// Fails with `signal.reason` so aborted calls reject and aborted streams throw like `fetch`: an `AbortError` by default.
 const abortEffect = (signal: AbortSignal | undefined) =>
  signal === undefined
    ? Effect.never
-    : Effect.callback<void>((resume) => {
+    : Effect.callback<never, unknown>((resume) => {
        if (signal.aborted) {
-          resume(Effect.void)
+          resume(Effect.fail(signal.reason))
          return
        }
-        const onAbort = () => resume(Effect.void)
+        const onAbort = () => resume(Effect.fail(signal.reason))
        signal.addEventListener("abort", onAbort, { once: true })
        return Effect.sync(() => signal.removeEventListener("abort", onAbort))
      })
@@ -68,14 +69,14 @@ export const make = (options: Options = {}) => {

  /** Run any package Effect (for example `LLMClient.compact(...)`) inside this runtime. */
  const run = <A, E>(effect: Effect.Effect<A, E, Services>, options?: RunOptions) =>
-    runtime.runPromise(effect, { signal: options?.signal })
+    runtime.runPromise(Effect.raceFirst(effect, abortEffect(options?.signal)))

  const iterate = <A, E>(stream: Stream.Stream<A, E, Services>, options?: RunOptions): AsyncIterable<A> =>
    Stream.toAsyncIterable(
      Stream.unwrap(
        runtime.contextEffect.pipe(
          Effect.map(
-            (context): Stream.Stream<A, E> =>
+            (context): Stream.Stream<A, unknown> =>
              stream.pipe(Stream.interruptWhen(abortEffect(options?.signal)), Stream.provideContext(context)),
          ),
        ),
@@ -3,6 +3,7 @@ import { Protocol } from "../route/protocol.js"
 import type { LanguageModelCompatibility } from "../schema/index.js"
 import { OpenAIChat } from "./openai-chat.js"
 import { JsonObject, ProviderShared } from "./shared.js"
+import { cacheControl } from "./utils/cache.js"
 import { OpenResponsesOptions } from "./utils/open-responses-options.js"

 export type ReasoningEffort = OpenResponsesOptions.ReasoningEffort
@@ -68,7 +69,7 @@ export const protocol = Protocol.make({
    from: Effect.fn("AlibabaChat.fromRequest")(function* (req) {
      const opts = yield* ProviderShared.validateWith(Schema.decodeUnknownEffect(Options))(req.providerOptions ?? {})
      return {
-        ...(yield* OpenAIChat.protocol.body.from(req)),
+        ...(yield* OpenAIChat.fromRequest(req, { cacheControl: cacheControl() })),
        enable_thinking: opts.enableThinking,
        // Alibaba also rejects an explicit budget that is not below `max_completion_tokens`.
        thinking_budget:
@@ -29,6 +29,7 @@ import { JsonObject, knownString, optionalArray, optionalNull, ProviderShared }
 import { classifyProviderFailure } from "../provider-error.js"
 import { effortUpdate, resolveEffortUpdates } from "../effort-updates.js"
 import * as Cache from "./utils/cache.js"
+import { claudeVersion, supportsThinkingBlockBinding, THINKING_BINDING_BETA } from "./utils/claude-model.js"
 import { Lifecycle } from "./utils/lifecycle.js"
 import { ToolStream } from "./utils/tool-stream.js"

@@ -450,6 +451,9 @@ const AnthropicStreamDelta = Schema.Struct({
  signature: Schema.optional(Schema.String),
  stop_reason: optionalNull(Schema.String),
  stop_sequence: optionalNull(Schema.String),
+  stop_details: optionalNull(
+    Schema.Struct({ category: optionalNull(Schema.String), explanation: optionalNull(Schema.String) }),
+  ),
 })
 type AnthropicStreamDelta = Schema.Schema.Type<typeof AnthropicStreamDelta>
 const decodeAnthropicStreamDelta = Schema.decodeUnknownOption(AnthropicStreamDelta)
@@ -803,15 +807,12 @@ const requireThinkingSignature = (request: LLMRequest) => {
 // Mid-conversation system messages became available with Opus 4.8 and version
 // 5 of the other supported Claude families. Treat later family versions as
 // compatible without assuming that every Anthropic Messages model is Claude.
+// Opus 4.8 and every Claude 5 model accept mid-conversation system messages; later versions inherit support.
 const supportsNativeSystemUpdates = (request: LLMRequest) => {
-  const match = /(?:^|[./])claude-(fable|haiku|mythos|opus|sonnet)-(\d+)(?:[.-](\d+))?/.exec(
-    String(request.model.id).toLowerCase(),
-  )
-  if (!match) return false
-  const major = Number(match[2])
-  if (match[1] !== "opus") return major >= 5
-  if (major !== 4) return major >= 5
-  return match[3] !== undefined && match[3].length <= 2 && Number(match[3]) >= 8
+  const version = claudeVersion(String(request.model.id))
+  if (version === undefined) return false
+  if (version.family === "opus" && version.major === 4) return version.minor >= 8
+  return version.major >= 5
 }

 const endsInServerToolUse = (message: LLMRequest["messages"][number]) => {
@@ -988,29 +989,13 @@ const lowerMessages = Effect.fn("AnthropicMessages.lowerMessages")(function* (
  return messages
 })

-// Accept gateway namespaces and Vertex suffixes without treating a snapshot date as a minor version.
-const claudeVersion = (id: string) => {
-  const match = /(?:^|[./])claude-(?<family>[a-z]+)-(?<major>\d+)(?:[.-](?<minor>\d{1,2}))?(?:$|[-:@])/.exec(
-    id.toLowerCase(),
-  )?.groups
-  if (!match) return undefined
-  return { family: match.family, major: Number(match.major), minor: Number(match.minor ?? 0) }
-}
-
-const supportsThinkingBlockBinding = (model: LLMRequest["model"]) => {
-  const override = model.compatibility?.supportsThinkingBlockBinding
-  if (override !== undefined) return override
-  const version = claudeVersion(model.id)
-  return version !== undefined && (version.major > 5 || (version.major === 5 && version.minor >= 1))
-}
-
+// Per-turn effort started with Claude Opus 5 and every Claude 5.1 model; later versions of any family inherit it.
 const supportsEffortUpdates = (model: LLMRequest["model"]) => {
  const override = model.compatibility?.supportsEffortUpdates
  if (override !== undefined) return override
  const version = claudeVersion(model.id)
  if (version === undefined) return false
-  if (version.family === "opus") return version.major >= 5
-  if (version.family !== "fable" && version.family !== "mythos") return false
+  if (version.family === "opus" && version.major >= 5) return true
  return version.major > 5 || (version.major === 5 && version.minor >= 1)
 }

@@ -1432,10 +1417,14 @@ const onMessageDelta = (
      stopSequence === null || stopSequence === undefined
        ? state.pendingFinish?.providerMetadata
        : providerMetadata(state.providerMetadataKey, { stopSequence })
+    const category = event.delta?.stop_details?.category
+    const explanation = event.delta?.stop_details?.explanation
    return {
      reason: {
        normalized: mapFinishReason(stopReason),
        raw: stopReason,
+        ...(category ? { category } : {}),
+        ...(explanation ? { explanation } : {}),
      },
      providerMetadata: finishMetadata,
    }
@@ -1652,8 +1641,7 @@ function requiredBetaHeaders(body: Pick<AnthropicMessagesBody, "messages" | "con
    betas.push("mid-conversation-output-config-2026-07-01")

  const thinking = body.thinking
-  if (thinking && thinking.type !== "disabled" && thinking.block_binding)
-    betas.push("thinking-binding-controls-2026-08-01")
+  if (thinking && thinking.type !== "disabled" && thinking.block_binding) betas.push(THINKING_BINDING_BETA)
  return betas
 }

@@ -110,8 +110,11 @@ const fromRequest = Effect.fn("AssemblyAITranscription.fromRequest")(function* (
        language_code: request.language,
        language_detection: request.language === undefined ? true : undefined,
        prompt: request.prompt,
-        // Turn-level `utterances`, the only segments AssemblyAI returns, require speaker labels.
-        speaker_labels: request.diarize === true || request.timestamps === "segment" ? true : undefined,
+        // Turn-level `utterances`, the only segments AssemblyAI returns, and `speakers_expected` require speaker labels.
+        speaker_labels:
+          request.diarize === true || request.timestamps === "segment" || request.speakers !== undefined
+            ? true
+            : undefined,
        speakers_expected: request.speakers,
      },
      request.providerOptions,
@@ -155,8 +158,7 @@ const decodeResult = Effect.fn("AssemblyAITranscription.decodeResult")(function*
  const error = transcript.error ?? undefined
  if (status === "failed")
    return yield* output.ended("failed", `${route.name} transcription failed${error === undefined ? "" : `: ${error}`}`)
-  if (status !== "completed")
-    return yield* output.invalid(`${route.name} transcript ${context.token.transcriptID} has not finished`)
+  if (status !== "completed") return yield* output.pending(context.token.transcriptID)
  const duration = transcript.audio_duration ?? undefined
  return new TranscriptionResponse({
    text: transcript.text ?? "",
@@ -23,6 +23,7 @@ import { JsonObject, optionalArray, ProviderShared } from "./shared.js"
 import { BedrockAuth } from "./utils/bedrock-auth.js"
 import { BedrockCache } from "./utils/bedrock-cache.js"
 import { BedrockMedia } from "./utils/bedrock-media.js"
+import { supportsThinkingBlockBinding, THINKING_BINDING_BETA } from "./utils/claude-model.js"
 import { Lifecycle } from "./utils/lifecycle.js"
 import { MistralToolID } from "./utils/mistral-tool-id.js"
 import { ToolStream } from "./utils/tool-stream.js"
@@ -318,6 +319,9 @@ const lowerToolResult = Effect.fn("BedrockConverse.lowerToolResult")(function* (
  } satisfies BedrockToolResultBlock
 })

+// Keep Claude and Nova tool-result images inline; put other models' images beside the result.
+const keepToolImagesInline = (id: string) => id.includes("anthropic.claude-") || id.includes("amazon.nova-")
+
 const lowerMessages = Effect.fn("BedrockConverse.lowerMessages")(function* (
  request: LLMRequest,
  breakpoints: BedrockCache.Breakpoints,
@@ -327,8 +331,19 @@ const lowerMessages = Effect.fn("BedrockConverse.lowerMessages")(function* (
  // Mistral can reject replay IDs even when they satisfy Converse's broader ID syntax.
  const normalizeID = request.model.id.includes("mistral.") ? MistralToolID.normalizer(request) : (id: string) => id
  const providerMetadataKey = request.model.route.providerMetadataKey ?? String(request.model.provider)
+  const hoistImages = !keepToolImagesInline(request.model.id)
+  // Bedrock expects parallel tool results before any images hoisted beside them.
+  const pendingImages: BedrockMedia.ImageBlock[] = []
+  const flushImages = () => {
+    if (pendingImages.length === 0) return
+    const previous = messages.at(-1)
+    if (previous?.role === "user")
+      messages[messages.length - 1] = { role: "user", content: [...previous.content, ...pendingImages] }
+    pendingImages.length = 0
+  }

  for (const message of request.messages) {
+    if (message.role !== "tool") flushImages()
    if (message.role === "system") {
      const part = yield* ProviderShared.wrappedSystemUpdate("Bedrock Converse", message)
      const content = textWithCache(breakpoints, part.text, part.cache)
@@ -402,7 +417,22 @@ const lowerMessages = Effect.fn("BedrockConverse.lowerMessages")(function* (
    for (const part of message.content) {
      if (!ProviderShared.supportsContent(part, ["tool-result"]))
        return yield* ProviderShared.unsupportedContent("Bedrock Converse", "tool", ["tool-result"])
-      content.push(yield* lowerToolResult(part, documentNames, normalizeID))
+      const result = yield* lowerToolResult(part, documentNames, normalizeID)
+      const images: BedrockMedia.ImageBlock[] = hoistImages
+        ? result.toolResult.content.filter((item) => "image" in item)
+        : []
+      const nonImageContent = result.toolResult.content.filter((item) => !("image" in item))
+      content.push(
+        images.length === 0
+          ? result
+          : {
+              toolResult: {
+                ...result.toolResult,
+                content: nonImageContent.length > 0 ? nonImageContent : [{ text: "See attached image." }],
+              },
+            },
+      )
+      pendingImages.push(...images)
      const cachePoint = BedrockCache.block(breakpoints, part.cache)
      if (cachePoint) content.push(cachePoint)
    }
@@ -412,6 +442,7 @@ const lowerMessages = Effect.fn("BedrockConverse.lowerMessages")(function* (
    else messages.push({ role: "user", content })
  }

+  flushImages()
  return messages
 })

@@ -443,6 +474,23 @@ const decodeOptions = ProviderShared.validateWith(Schema.decodeUnknownEffect(Opt
 // Claude on Bedrock requires the thinking budget below `maxTokens`, with a minimum of 1,024.
 const MIN_THINKING_BUDGET = 1_024

+const isThinkingDisabled = Schema.is(
+  Schema.Struct({
+    additionalModelRequestFields: Schema.Struct({ thinking: Schema.Struct({ type: Schema.Literal("disabled") }) }),
+  }),
+)
+
+// Claude 5.1+ binds each thinking signature to the prefix above it. Ask Bedrock to drop the affected blocks instead of
+// failing when that prefix changes. `http.body` overlays this field by field, so callers can still override it.
+const applyThinkingBindingDefault = (request: LLMRequest, thinking: Readonly<Record<string, unknown>> | undefined) => {
+  if (isThinkingDisabled(request.http?.body)) return thinking
+  if (!supportsThinkingBlockBinding(request.model)) return thinking
+  return {
+    ...(thinking ?? { type: "adaptive" as const }),
+    block_binding: { prefix_mismatch_behavior: "drop_block" },
+  }
+}
+
 const fromRequest = Effect.fn("BedrockConverse.fromRequest")(function* (request: LLMRequest) {
  const toolChoice = request.toolChoice ? yield* lowerToolChoice(request.toolChoice) : undefined
  const flattened = ProviderShared.flattenToolRequest(request)
@@ -450,7 +498,8 @@ const fromRequest = Effect.fn("BedrockConverse.fromRequest")(function* (request:
  const options = yield* decodeOptions(request.providerOptions ?? {})
  const maxTokens =
    isNova2(request.model) && isHighReasoningEffort(request.http?.body) ? undefined : generation?.maxTokens
-  const thinking =
+  const thinking = applyThinkingBindingDefault(
+    request,
    options.thinking === undefined
      ? undefined
      : {
@@ -460,7 +509,8 @@ const fromRequest = Effect.fn("BedrockConverse.fromRequest")(function* (request:
            maxTokens,
            MIN_THINKING_BUDGET,
          ),
-        }
+        },
+  )
  // Bedrock-Claude shares Anthropic's 4-breakpoint cap. Spend the budget in
  // tools → system → messages order to favour the highest-impact prefixes.
  const breakpoints = BedrockCache.breakpoints(request.model.id)
@@ -509,6 +559,8 @@ const fromRequest = Effect.fn("BedrockConverse.fromRequest")(function* (request:
        : {
            ...(generation?.topK === undefined ? {} : { top_k: generation.topK }),
            ...(thinking === undefined ? {} : { thinking }),
+            // Converse takes Anthropic betas in the body, and Bedrock rejects `block_binding` without this one.
+            ...(thinking?.block_binding === undefined ? {} : { anthropic_beta: [THINKING_BINDING_BETA] }),
          },
  }
 })
@@ -555,7 +607,7 @@ interface ParserState {
  readonly hasToolCalls: boolean
  readonly lifecycle: Lifecycle.State
  readonly reasoningSignatures: Readonly<Record<number, string>>
-  readonly reasoningRedactedContent: Readonly<Record<number, ReadonlyArray<Uint8Array>>>
+  readonly reasoningRedactedContent: Readonly<Record<number, Uint8Array[]>>
 }

 const encodeRedactedContent = (chunks: ReadonlyArray<Uint8Array>) => Encoding.encodeBase64(concatBytes(chunks))
@@ -605,10 +657,9 @@ const step = (state: ParserState, event: BedrockEvent) =>
      const index = event.contentBlockDelta.contentBlockIndex
      const reasoning = event.contentBlockDelta.delta.reasoningContent
      const events: LLMEvent[] = []
-      const redactedChunks = yield* (() => {
+      const redactedChunk = yield* (() => {
        if (reasoning.redactedContent === undefined) return Effect.succeed(undefined)
        return Effect.fromResult(Encoding.decodeBase64(reasoning.redactedContent)).pipe(
-          Effect.map((chunk) => [...(state.reasoningRedactedContent[index] ?? []), chunk]),
          Effect.mapError((cause) =>
            ProviderShared.eventError(
              ADAPTER,
@@ -619,17 +670,21 @@ const step = (state: ParserState, event: BedrockEvent) =>
          ),
        )
      })()
-      const redactedData = redactedChunks === undefined ? reasoning.data : encodeRedactedContent(redactedChunks)
+      const redactedChunks = state.reasoningRedactedContent[index] ?? []
+      if (redactedChunk !== undefined) redactedChunks.push(redactedChunk)
      const metadata = (() => {
        if (reasoning.signature) return providerMetadata(state.providerMetadataKey, { signature: reasoning.signature })
-        if (redactedData !== undefined) return providerMetadata(state.providerMetadataKey, { redactedData })
+        if (redactedChunk === undefined && reasoning.data !== undefined)
+          return providerMetadata(state.providerMetadataKey, { redactedData: reasoning.data })
      })()
      const lifecycle = (() => {
-        if (reasoning.text === undefined && metadata === undefined) return state.lifecycle
-        return Lifecycle.reasoningDelta(state.lifecycle, events, `reasoning-${index}`, reasoning.text ?? "", metadata)
+        if (reasoning.text !== undefined || metadata !== undefined)
+          return Lifecycle.reasoningDelta(state.lifecycle, events, `reasoning-${index}`, reasoning.text ?? "", metadata)
+        if (redactedChunk !== undefined) return Lifecycle.reasoningStart(state.lifecycle, events, `reasoning-${index}`)
+        return state.lifecycle
      })()
      const reasoningRedactedContent = (() => {
-        if (redactedChunks !== undefined) return { ...state.reasoningRedactedContent, [index]: redactedChunks }
+        if (redactedChunk !== undefined) return { ...state.reasoningRedactedContent, [index]: redactedChunks }
        if (reasoning.data === undefined) return state.reasoningRedactedContent
        return Object.fromEntries(
          Object.entries(state.reasoningRedactedContent).filter(([key]) => key !== String(index)),
@@ -765,7 +820,19 @@ const onHalt = (state: ParserState): ReadonlyArray<LLMEvent> => {
    return state.finishReason.normalized
  })()
  const events: LLMEvent[] = []
-  Lifecycle.finish(state.lifecycle, events, {
+  const lifecycle = Object.entries(state.reasoningRedactedContent).reduce((current, [index, chunks]) => {
+    const signature = state.reasoningSignatures[Number(index)]
+    return Lifecycle.reasoningEnd(
+      current,
+      events,
+      `reasoning-${index}`,
+      providerMetadata(
+        state.providerMetadataKey,
+        signature ? { signature } : { redactedData: encodeRedactedContent(chunks) },
+      ),
+    )
+  }, state.lifecycle)
+  Lifecycle.finish(lifecycle, events, {
    reason: {
      ...state.finishReason,
      normalized,
@@ -31,13 +31,21 @@ export type Request = ImageRequestFor<BlackForestLabsImageOptions>
 // 2. Token and response schemas
 // ---------------------------------------------------------------------------

-/** Regional clusters answer on different hosts, so the returned `polling_url` is followed verbatim. */
-export const Token = Schema.Struct({ id: Schema.String, pollingURL: Schema.String })
+/**
+ * Regional clusters answer on different hosts, so the returned `polling_url` is followed verbatim. BFL reports the
+ * credit cost on submit, so it rides on the token; it is optional so tokens persisted before it existed still decode.
+ */
+export const Token = Schema.Struct({
+  id: Schema.String,
+  pollingURL: Schema.String,
+  cost: Schema.optionalKey(Schema.Number),
+})
 export type Token = Schema.Schema.Type<typeof Token>

 const StartResponse = Schema.Struct({
  id: Schema.String,
  polling_url: Schema.String,
+  cost: optionalNull(Schema.Number),
 })

 const Result = Schema.Struct({
@@ -145,7 +153,11 @@ const fromRequest = Effect.fn("BlackForestLabsImages.fromRequest")(function* (re
 // ---------------------------------------------------------------------------

 const decodeStart = route.decodeStarted(StartResponse, (value) => ({
-  token: { id: value.id, pollingURL: value.polling_url },
+  token: {
+    id: value.id,
+    pollingURL: value.polling_url,
+    ...(value.cost === undefined || value.cost === null ? {} : { cost: value.cost }),
+  },
  snapshot: { id: value.id, status: "queued" },
 }))

@@ -169,14 +181,16 @@ const decodeResult = Effect.fn("BlackForestLabsImages.decodeResult")(function* (
  if (isModerated(document.status)) return yield* output.contentPolicy(`${route.name} moderated the generation`)
  if (status === "failed" || status === "expired")
    return yield* output.ended(status, `${route.name} generation ${context.token.id} ended with ${document.status}`)
-  if (status !== "completed" || document.result === undefined || document.result === null)
+  if (status !== "completed") return yield* output.pending(context.token.id)
+  if (document.result === undefined || document.result === null)
    return yield* output.invalid(`${route.name} generation ${context.token.id} has no result`)
  const { sample, seed, prompt, ...rest } = document.result
+  // A settled `cost` on the result supersedes the submit-time cost carried on the token.
+  const cost = document.cost ?? context.token.cost
  return new ImageResponse({
    // `sample` is a signed URL that expires 10 minutes after the result is ready, so it is downloaded now.
    images: [yield* context.materialize(Media.url(sample))],
-    usage:
-      document.cost === undefined || document.cost === null ? undefined : { type: "credits", credits: document.cost },
+    usage: cost === undefined ? undefined : { type: "credits", credits: cost },
    providerMetadata: {
      bfl: { id: context.token.id, seed: seed ?? undefined, prompt: prompt ?? undefined, ...rest },
    },
@@ -0,0 +1,326 @@
+import { Effect, Schema } from "effect"
+import { Route } from "../route/client.js"
+import { Endpoint } from "../route/endpoint.js"
+import { Framing } from "../route/framing.js"
+import { Protocol } from "../route/protocol.js"
+import { LLMEvent, Usage, type FinishReasonDetails, type LLMRequest } from "../schema/index.js"
+import { ProviderShared } from "./shared.js"
+import { Lifecycle } from "./utils/lifecycle.js"
+import { ToolStream } from "./utils/tool-stream.js"
+
+const ADAPTER = "cohere-chat"
+export const DEFAULT_BASE_URL = "https://api.cohere.com/v2"
+
+const Options = Schema.Struct({
+  thinking: Schema.optional(
+    Schema.Struct({
+      type: Schema.optional(Schema.Literals(["enabled", "disabled"])),
+      tokenBudget: Schema.optional(Schema.Int.check(Schema.isGreaterThan(0))),
+    }),
+  ),
+})
+export type ProviderOptionsInput = Schema.Schema.Type<typeof Options>
+
+const Content = Schema.Union([
+  Schema.Struct({ type: Schema.Literal("text"), text: Schema.String }),
+  Schema.Struct({ type: Schema.Literal("thinking"), thinking: Schema.String }),
+  Schema.Struct({ type: Schema.Literal("image_url"), image_url: Schema.Struct({ url: Schema.String }) }),
+])
+const ToolCall = Schema.Struct({
+  id: Schema.String,
+  type: Schema.Literal("function"),
+  function: Schema.Struct({ name: Schema.String, arguments: Schema.String }),
+})
+const Message = Schema.Struct({
+  role: Schema.Literals(["system", "user", "assistant", "tool"]),
+  content: Schema.optional(Schema.Union([Schema.String, Schema.Array(Content)])),
+  tool_calls: Schema.optional(Schema.Array(ToolCall)),
+  tool_call_id: Schema.optional(Schema.String),
+  tool_plan: Schema.optional(Schema.String),
+})
+const Body = Schema.Struct({
+  model: Schema.String,
+  messages: Schema.Array(Message),
+  stream: Schema.Literal(true),
+  tools: Schema.optional(
+    Schema.Array(
+      Schema.Struct({
+        type: Schema.Literal("function"),
+        function: Schema.Struct({
+          name: Schema.String,
+          description: Schema.optional(Schema.String),
+          parameters: Schema.Unknown,
+        }),
+      }),
+    ),
+  ),
+  tool_choice: Schema.optional(Schema.Literals(["NONE", "REQUIRED"])),
+  thinking: Schema.optional(Schema.Struct({ type: Schema.String, token_budget: Schema.optional(Schema.Number) })),
+  max_tokens: Schema.optional(Schema.Number),
+  temperature: Schema.optional(Schema.Number),
+  p: Schema.optional(Schema.Number),
+  k: Schema.optional(Schema.Number),
+  seed: Schema.optional(Schema.Number),
+  stop_sequences: Schema.optional(Schema.Array(Schema.String)),
+  frequency_penalty: Schema.optional(Schema.Number),
+  presence_penalty: Schema.optional(Schema.Number),
+})
+const TokenCounts = Schema.Struct({
+  input_tokens: Schema.optional(Schema.Number),
+  output_tokens: Schema.optional(Schema.Number),
+  reasoning_tokens: Schema.optional(Schema.Number),
+})
+const NativeUsage = Schema.Struct({
+  tokens: Schema.optional(TokenCounts),
+  billed_units: Schema.optional(TokenCounts),
+  cached_tokens: Schema.optional(Schema.Number),
+})
+const Event = Schema.Union([
+  Schema.Struct({ type: Schema.Literal("message-start") }),
+  Schema.Struct({
+    type: Schema.Literals(["content-start", "content-delta"]),
+    index: Schema.Number,
+    delta: Schema.Struct({
+      message: Schema.Struct({
+        content: Schema.Struct({ text: Schema.optional(Schema.String), thinking: Schema.optional(Schema.String) }),
+      }),
+    }),
+  }),
+  Schema.Struct({ type: Schema.Literal("content-end"), index: Schema.Number }),
+  Schema.Struct({
+    type: Schema.Literal("tool-plan-delta"),
+    delta: Schema.Struct({ message: Schema.Struct({ tool_plan: Schema.String }) }),
+  }),
+  Schema.Struct({
+    type: Schema.Literals(["tool-call-start", "tool-call-delta"]),
+    index: Schema.Number,
+    delta: Schema.Struct({
+      message: Schema.Struct({
+        tool_calls: Schema.Struct({
+          id: Schema.optional(Schema.String),
+          function: Schema.Struct({ name: Schema.optional(Schema.String), arguments: Schema.optional(Schema.String) }),
+        }),
+      }),
+    }),
+  }),
+  Schema.Struct({ type: Schema.Literal("tool-call-end"), index: Schema.Number }),
+  Schema.Struct({
+    type: Schema.Literal("message-end"),
+    delta: Schema.Struct({ finish_reason: Schema.String, usage: Schema.optional(NativeUsage) }),
+  }),
+  // Citation output is outside this basic chat surface.
+  Schema.Struct({ type: Schema.Literals(["citation-start", "citation-end"]) }),
+])
+type Event = typeof Event.Type
+type State = {
+  readonly lifecycle: Lifecycle.State
+  readonly tools: ToolStream.State<number>
+  readonly finished: boolean
+}
+
+const TOOL_CHOICE = { auto: undefined, none: "NONE", required: "REQUIRED", tool: "REQUIRED" } as const
+
+const fromRequest = Effect.fn("CohereChat.fromRequest")(function* (request: LLMRequest) {
+  const options = yield* ProviderShared.validateWith(Schema.decodeUnknownEffect(Options))(request.providerOptions ?? {})
+  const flattened = ProviderShared.flattenToolRequest(request)
+  const messages: (typeof Message.Type)[] = request.system.length
+    ? [{ role: "system", content: ProviderShared.joinText(request.system) }]
+    : []
+  for (const message of flattened.request.messages) {
+    if (message.role === "system") {
+      messages.push({ role: "user", content: (yield* ProviderShared.wrappedSystemUpdate("Cohere Chat", message)).text })
+      continue
+    }
+    if (message.role === "tool") {
+      for (const part of message.content) {
+        if (part.type !== "tool-result")
+          return yield* ProviderShared.unsupportedContent("Cohere Chat", "tool", ["tool-result"])
+        if (part.result.type === "content" && part.result.value.some((item) => item.type === "file"))
+          return yield* ProviderShared.invalidRequest("Cohere Chat does not support file content in tool results")
+        messages.push({ role: "tool", tool_call_id: part.id, content: ProviderShared.toolResultText(part) })
+      }
+      continue
+    }
+    const content: (typeof Content.Type)[] = []
+    const calls: (typeof ToolCall.Type)[] = []
+    const plans: string[] = []
+    for (const part of message.content) {
+      if (part.type === "text") {
+        content.push({ type: "text", text: part.text })
+        continue
+      }
+      if (message.role === "assistant" && part.type === "reasoning") {
+        if (part.providerMetadata?.cohere?.toolPlan === true) plans.push(part.text)
+        else content.push({ type: "thinking", thinking: part.text })
+        continue
+      }
+      if (message.role === "assistant" && part.type === "tool-call") {
+        const args = ProviderShared.encodeJson(part.input)
+        calls.push({ id: part.id, type: "function", function: { name: part.name, arguments: args } })
+        continue
+      }
+      if (message.role === "user" && part.type === "media" && part.media.mediaType.startsWith("image/")) {
+        const url =
+          ProviderShared.mediaUrl(part.media) ??
+          (yield* ProviderShared.requireInlineMedia("Cohere Chat", part.media)).dataUrl
+        content.push({ type: "image_url", image_url: { url } })
+        continue
+      }
+      return yield* ProviderShared.unsupportedContent(
+        "Cohere Chat",
+        message.role,
+        message.role === "user" ? ["text", "media"] : ["text", "reasoning", "tool-call"],
+      )
+    }
+    messages.push({
+      role: message.role,
+      content: content.length ? content : undefined,
+      tool_calls: calls.length ? calls : undefined,
+      tool_plan: plans.length ? plans.join("") : undefined,
+    })
+  }
+  const selected = request.toolChoice?.type === "tool" ? request.toolChoice.name : undefined
+  const tools = selected === undefined ? flattened.tools : flattened.tools.filter((tool) => tool.name === selected)
+  if (selected !== undefined && tools.length === 0)
+    return yield* ProviderShared.invalidRequest("Cohere Chat tool choice must name an available tool")
+  if (tools.some((tool) => tool.native !== undefined))
+    return yield* ProviderShared.invalidRequest("Cohere Chat does not support provider-defined tools")
+  return {
+    model: request.model.id,
+    messages,
+    stream: true as const,
+    tools: tools.length
+      ? tools.map((tool) => ({
+          type: "function" as const,
+          function: { name: tool.name, description: tool.description, parameters: tool.inputSchema },
+        }))
+      : undefined,
+    tool_choice: TOOL_CHOICE[request.toolChoice?.type ?? "auto"],
+    thinking: options.thinking && {
+      type: options.thinking.type ?? "enabled",
+      // Cohere rejects budgets above max_tokens; fitting also leaves room for the answer.
+      token_budget:
+        options.thinking.tokenBudget === undefined
+          ? undefined
+          : ProviderShared.fitThinkingBudget(options.thinking.tokenBudget, request.generation?.maxTokens),
+    },
+    max_tokens: request.generation?.maxTokens,
+    temperature: request.generation?.temperature,
+    p: request.generation?.topP,
+    k: request.generation?.topK,
+    seed: request.generation?.seed,
+    stop_sequences: request.generation?.stop,
+    frequency_penalty: request.generation?.frequencyPenalty,
+    presence_penalty: request.generation?.presencePenalty,
+  }
+})
+
+const finishReason = (raw: string): FinishReasonDetails => {
+  switch (raw) {
+    case "COMPLETE":
+    case "STOP_SEQUENCE":
+      return { normalized: "stop", raw }
+    case "MAX_TOKENS":
+      return { normalized: "length", raw }
+    case "TOOL_CALL":
+      return { normalized: "tool-calls", raw }
+    case "ERROR":
+    case "TIMEOUT":
+      return { normalized: "error", raw }
+    default:
+      return { normalized: "unknown", raw }
+  }
+}
+
+const mapUsage = (usage: typeof NativeUsage.Type) =>
+  new Usage({
+    inputTokens: usage.tokens?.input_tokens,
+    outputTokens: usage.tokens?.output_tokens,
+    nonCachedInputTokens: ProviderShared.subtractTokens(usage.tokens?.input_tokens, usage.cached_tokens),
+    cacheReadInputTokens: usage.cached_tokens,
+    reasoningTokens: usage.tokens?.reasoning_tokens,
+    totalTokens: ProviderShared.totalTokens(usage.tokens?.input_tokens, usage.tokens?.output_tokens, undefined),
+    providerMetadata: { cohere: usage },
+  })
+
+// Lifecycle deltas open blocks on demand and ends are no-ops for closed blocks, so content-start needs no handling.
+const step = Effect.fn("CohereChat.step")(function* (state: State, event: Event) {
+  const events: LLMEvent[] = []
+  switch (event.type) {
+    case "message-start":
+      return [{ ...state, lifecycle: Lifecycle.stepStart(state.lifecycle, events) }, events] as const
+    case "content-delta": {
+      const id = String(event.index)
+      const content = event.delta.message.content
+      const lifecycle =
+        content.thinking !== undefined
+          ? Lifecycle.reasoningDelta(state.lifecycle, events, id, content.thinking)
+          : Lifecycle.textDelta(state.lifecycle, events, id, content.text ?? "")
+      return [{ ...state, lifecycle }, events] as const
+    }
+    case "content-end": {
+      const id = String(event.index)
+      const lifecycle = Lifecycle.textEnd(Lifecycle.reasoningEnd(state.lifecycle, events, id), events, id)
+      return [{ ...state, lifecycle }, events] as const
+    }
+    case "tool-plan-delta": {
+      const plan = event.delta.message.tool_plan
+      const lifecycle = Lifecycle.reasoningDelta(state.lifecycle, events, "tool-plan", plan, {
+        cohere: { toolPlan: true },
+      })
+      return [{ ...state, lifecycle }, events] as const
+    }
+    case "tool-call-start":
+    case "tool-call-delta": {
+      const call = event.delta.message.tool_calls
+      const result = ToolStream.appendOrStart(
+        ADAPTER,
+        state.tools,
+        event.index,
+        { id: call.id, name: call.function.name, text: call.function.arguments ?? "" },
+        "Cohere tool call is missing id or name",
+      )
+      if (ToolStream.isError(result)) return yield* result
+      return [{ ...state, tools: result.tools }, result.events] as const
+    }
+    case "tool-call-end": {
+      const result = yield* ToolStream.finish(ADAPTER, state.tools, event.index)
+      return [{ ...state, tools: result.tools }, result.events ?? []] as const
+    }
+    case "message-end": {
+      const pending = yield* ToolStream.finishAll(ADAPTER, state.tools)
+      events.push(...pending.events)
+      const lifecycle = Lifecycle.finish(state.lifecycle, events, {
+        reason: finishReason(event.delta.finish_reason),
+        usage: event.delta.usage && mapUsage(event.delta.usage),
+      })
+      return [{ tools: pending.tools, lifecycle, finished: true }, events] as const
+    }
+    default:
+      return [state, events] as const
+  }
+})
+
+export const protocol = Protocol.make({
+  id: ADAPTER,
+  body: { schema: Body, from: fromRequest },
+  stream: {
+    event: Protocol.jsonEvent(Event),
+    initial: (): State => ({ lifecycle: Lifecycle.initial(), tools: ToolStream.empty(), finished: false }),
+    step,
+    terminal: (event) => event.type === "message-end",
+    onHalt: (state) =>
+      state.finished
+        ? Effect.succeed([])
+        : Effect.fail(ProviderShared.eventError(ADAPTER, "Cohere stream ended without message-end")),
+  },
+})
+export const route = Route.make({
+  id: ADAPTER,
+  provider: "cohere",
+  providerMetadataKey: "cohere",
+  protocol,
+  endpoint: Endpoint.path("/chat", { baseURL: DEFAULT_BASE_URL }),
+  framing: Framing.sse,
+})
+export * as CohereChat from "./cohere-chat.js"
@@ -67,6 +67,9 @@ const queryParameters = (request: Request) => {
 }

 const fromRequest = Effect.fn("DeepgramSpeech.fromRequest")(function* (request: Request) {
+  // Not in `unsupported`: that list would also reject `timestamps: false`, which asks for nothing.
+  if (request.timestamps === true)
+    return yield* route.unsupported("media.timestamps", `${route.name} does not return timestamps`)
  if (
    request.format !== undefined &&
    FORMATS[request.format] === undefined &&
@@ -86,24 +89,32 @@ const fromRequest = Effect.fn("DeepgramSpeech.fromRequest")(function* (request:
 // 6. Stream parsing
 // ---------------------------------------------------------------------------

-const HEADERLESS_ENCODINGS: Readonly<Record<string, SpeechStream.PcmEncoding>> = {
-  linear16: "pcm_s16le",
-  mulaw: "pcm_mulaw",
-  alaw: "pcm_alaw",
+/** Deepgram wraps raw encodings in WAV unless `container` is `none`, and defaults their sample rate per encoding. */
+const HEADERLESS_ENCODINGS: Readonly<
+  Record<string, { readonly encoding: SpeechStream.PcmEncoding; readonly sampleRate: number }>
+> = {
+  linear16: { encoding: "pcm_s16le", sampleRate: 24000 },
+  mulaw: { encoding: "pcm_mulaw", sampleRate: 8000 },
+  alaw: { encoding: "pcm_alaw", sampleRate: 8000 },
 }

 const finish = (state: State, context: MediaProtocol.ResponseContext<Request>) => {
  const headers = context.http.headers
  const mediaType = headers["content-type"]
  const format = audioFormat(context.request)
-  const encoding = HEADERLESS_ENCODINGS[format.encoding ?? ""]
+  const headerless = HEADERLESS_ENCODINGS[format.encoding ?? ""]
+  const container = format.container ?? (headerless === undefined ? undefined : "wav")
  const requestID = headers["dg-request-id"]
  const modelName = headers["dg-model-name"]
  return SpeechStream.finish(route, state, {
-    ...(format.container === "none" && encoding !== undefined
-      ? SpeechStream.pcm(encoding, SpeechStream.sampleRate(mediaType), mediaType)
+    ...(container === "none" && headerless !== undefined
+      ? SpeechStream.pcm(
+          headerless.encoding,
+          SpeechStream.sampleRate(mediaType) ?? context.request.providerOptions?.sampleRate ?? headerless.sampleRate,
+          mediaType,
+        )
      : // Deepgram's default encoding is MP3; WAV is a container around any encoding.
-        { mediaType, info: { format: format.container === "wav" ? "wav" : (format.encoding ?? "mp3") } }),
+        { mediaType, info: { format: container === "wav" ? "wav" : (format.encoding ?? "mp3") } }),
    usage: SpeechStream.headerUsage("characters", headers["dg-char-count"]),
    providerMetadata:
      requestID === undefined && modelName === undefined
@@ -117,7 +128,7 @@ const finish = (state: State, context: MediaProtocol.ResponseContext<Request>) =
 // ---------------------------------------------------------------------------

 export const protocol = MediaProtocol.stream<Request, SpeechEvent, Uint8Array, State>(route, {
-  unsupported: ["voice", "language", "instructions", "timestamps"],
+  unsupported: ["voice", "language", "instructions"],
  body: { from: fromRequest },
  frames: (bytes) => bytes,
  initial: () => ({ chunks: [] }),
@@ -6,6 +6,7 @@ import { mergeJsonRecords, type OpenString } from "../schema/index.js"
 import { TranscriptionModel, TranscriptionResponse, type TranscriptionRequestFor } from "../transcription.js"
 import { ProviderShared } from "./shared.js"
 import { MediaInput } from "./utils/media-input.js"
+import { SpeakerTurns } from "./utils/speaker-turns.js"

 const route = MediaProtocol.identity({ id: "deepgram-transcription", name: "Deepgram", provider: "deepgram" })
 export const DEFAULT_BASE_URL = "https://api.deepgram.com"
@@ -115,16 +116,6 @@ const speaker = (value: number | undefined) => (value === undefined ? undefined

 const wordText = (word: typeof Word.Type) => word.punctuated_word ?? word.word

-// Utterances split on pauses, not speakers: the v2 diarizer labels a whole utterance with one speaker even when its
-// words change speaker, so segments split each utterance at speaker changes.
-const speakerTurns = (words: ReadonlyArray<typeof Word.Type>) =>
-  words.reduce<Array<Array<typeof Word.Type>>>((turns, word) => {
-    const last = turns.at(-1)
-    if (last === undefined || last[0].speaker !== word.speaker) return [...turns, [word]]
-    last.push(word)
-    return turns
-  }, [])
-
 const decodeResponse = Effect.fn("DeepgramTranscription.decodeResponse")(function* (
  response: HttpClientResponse.HttpClientResponse,
 ) {
@@ -136,6 +127,8 @@ const decodeResponse = Effect.fn("DeepgramTranscription.decodeResponse")(functio
  const requestID = output.value.metadata?.request_id
  return new TranscriptionResponse({
    text: alternative.transcript,
+    // Utterances split on pauses, not speakers: the v2 diarizer labels a whole utterance with one speaker even when
+    // its words change speaker, so segments split each utterance at speaker changes.
    segments: output.value.results.utterances?.flatMap((utterance) =>
      utterance.words === undefined || utterance.words.length === 0
        ? [
@@ -146,7 +139,7 @@ const decodeResponse = Effect.fn("DeepgramTranscription.decodeResponse")(functio
              speaker: speaker(utterance.speaker),
            },
          ]
-        : speakerTurns(utterance.words).map((turn) => ({
+        : SpeakerTurns.group(utterance.words, (word) => word.speaker).map((turn) => ({
            text: turn.map(wordText).join(" "),
            startSeconds: turn[0].start,
            endSeconds: turn[turn.length - 1].end,
@@ -0,0 +1,211 @@
+import { Effect, Schema } from "effect"
+import type { HttpClientResponse } from "effect/unstable/http"
+import { MediaProtocol } from "../route/media-protocol.js"
+import { MediaRoute } from "../route/media.js"
+import { mergeJsonRecords, type OpenString } from "../schema/index.js"
+import { TranscriptionModel, TranscriptionResponse, type TranscriptionRequestFor } from "../transcription.js"
+import { mediaTypeExtension } from "../utils/media-type.js"
+import { ProviderShared, optionalNull } from "./shared.js"
+import { MediaInput } from "./utils/media-input.js"
+import { SpeakerTurns } from "./utils/speaker-turns.js"
+
+const route = MediaProtocol.identity({
+  id: "elevenlabs-transcription",
+  name: "ElevenLabs Transcription",
+  provider: "elevenlabs",
+})
+export const DEFAULT_BASE_URL = "https://api.elevenlabs.io"
+export const PATH = "/v1/speech-to-text"
+
+// ---------------------------------------------------------------------------
+// 1. Public model input
+// ---------------------------------------------------------------------------
+
+export type ElevenLabsTranscriptionOptions = {
+  readonly tag_audio_events?: boolean
+  readonly timestamps_granularity?: OpenString<"none" | "word" | "character">
+  readonly diarization_threshold?: number
+  readonly file_format?: OpenString<"pcm_s16le_16" | "other">
+  readonly temperature?: number
+  readonly seed?: number
+  readonly keyterms?: ReadonlyArray<string>
+  readonly no_verbatim?: boolean
+  readonly detect_speaker_roles?: boolean
+  readonly use_speaker_library?: boolean
+  readonly entity_detection?: string | ReadonlyArray<string>
+  readonly entity_redaction?: string | ReadonlyArray<string>
+  readonly entity_redaction_mode?: OpenString<"redacted" | "entity_type" | "enumerated_entity_type">
+} & Record<string, unknown>
+
+export type Request = TranscriptionRequestFor<ElevenLabsTranscriptionOptions>
+
+// ---------------------------------------------------------------------------
+// 2. Response schema
+// ---------------------------------------------------------------------------
+
+/** `type` is `word`, `spacing` (the whitespace between words), or `audio_event` (`(laughter)`). */
+const Token = Schema.Struct({
+  text: Schema.String,
+  type: Schema.String,
+  start: optionalNull(Schema.Number),
+  end: optionalNull(Schema.Number),
+  speaker_id: optionalNull(Schema.String),
+  logprob: optionalNull(Schema.Number),
+})
+type Token = Schema.Schema.Type<typeof Token>
+
+const Transcript = Schema.Struct({
+  language_code: optionalNull(Schema.String),
+  text: Schema.String,
+  words: optionalNull(Schema.Array(Token)),
+  transcription_id: optionalNull(Schema.String),
+  audio_duration_secs: optionalNull(Schema.Number),
+})
+
+// ---------------------------------------------------------------------------
+// 5. Request body construction
+// ---------------------------------------------------------------------------
+
+/** Speaker turns are the only segments ElevenLabs can produce, and `num_speakers` only applies to diarization. */
+const diarizes = (request: Request) =>
+  request.diarize === true || request.timestamps === "segment" || request.speakers !== undefined
+
+const RESERVED_FORM_FIELDS = new Set([
+  "file",
+  "cloud_storage_url",
+  "source_url",
+  "model_id",
+  "language_code",
+  "diarize",
+  "num_speakers",
+])
+
+const validate = (request: Request, overlay: Record<string, unknown>) => {
+  // Webhook requests return 202 with no transcript; the result arrives at a configured webhook instead.
+  if (overlay.webhook === true)
+    return Effect.fail(route.unsupported("transcription.webhook", `${route.name} does not deliver to webhooks`))
+  // Separate multichannel output replaces the transcript with one transcript per channel.
+  if (overlay.use_multi_channel === true && overlay.multichannel_output_style !== "combined")
+    return Effect.fail(
+      route.unsupported(
+        "transcription.multichannel",
+        `${route.name} returns a single transcript; set multichannel_output_style: "combined" to merge channels`,
+      ),
+    )
+  if (overlay.timestamps_granularity === "none" && (request.timestamps === "word" || diarizes(request)))
+    return Effect.fail(
+      route.unsupported(
+        "media.timestamps",
+        `${route.name} cannot return word timestamps or speaker turns with timestamps_granularity: "none"`,
+      ),
+    )
+  return Effect.void
+}
+
+const fromRequest = Effect.fn("ElevenLabsTranscription.fromRequest")(function* (request: Request) {
+  const overlay = mergeJsonRecords(request.providerOptions, request.http?.body) ?? {}
+  yield* validate(request, overlay)
+  const form = new FormData()
+  const url = ProviderShared.mediaUrl(request.audio)
+  if (url === undefined) {
+    const extension = mediaTypeExtension(request.audio.mediaType)
+    const audio = yield* MediaInput.inlineBytes(route.id, request.audio)
+    form.append(
+      "file",
+      MediaInput.blob(audio, request.audio.mediaType),
+      extension === undefined ? "audio" : `audio.${extension}`,
+    )
+  }
+  MediaInput.appendFields(
+    form,
+    {
+      model_id: request.model.id,
+      // `cloud_storage_url` is deprecated in favor of `source_url`, which accepts any hosted audio or video URL.
+      source_url: url,
+      language_code: request.language,
+      diarize: diarizes(request) ? true : undefined,
+      num_speakers: request.speakers,
+    },
+    { overlay, reserved: RESERVED_FORM_FIELDS, repeatArrays: "key" },
+  )
+  return MediaProtocol.multipart(form)
+})
+
+// ---------------------------------------------------------------------------
+// 6. Response decoding
+// ---------------------------------------------------------------------------
+
+const decodeTranscript = route.decodeJson(Transcript)
+
+type TimedWord = Token & { readonly start: number; readonly end: number }
+
+const isTimedWord = (token: Token): token is TimedWord =>
+  token.type === "word" && typeof token.start === "number" && typeof token.end === "number"
+
+/** Turn text keeps the provider's own spacing tokens, so languages written without spaces are not re-spaced. */
+const speakerTurns = (tokens: ReadonlyArray<Token>) =>
+  SpeakerTurns.group(
+    tokens.filter((token) => token.type === "word" || token.type === "spacing"),
+    (token) => token.speaker_id,
+  ).flatMap((turn) => {
+    const words = turn.filter(isTimedWord)
+    if (words.length === 0) return []
+    return [
+      {
+        text: turn
+          .map((token) => token.text)
+          .join("")
+          .trim(),
+        startSeconds: words[0].start,
+        endSeconds: words[words.length - 1].end,
+        speaker: turn[0].speaker_id ?? undefined,
+      },
+    ]
+  })
+
+const decodeResponse = Effect.fn("ElevenLabsTranscription.decodeResponse")(function* (
+  response: HttpClientResponse.HttpClientResponse,
+  context: MediaProtocol.DecodeContext<Request>,
+) {
+  const output = yield* decodeTranscript(response)
+  const transcript = output.value
+  const tokens = transcript.words ?? []
+  const duration = transcript.audio_duration_secs ?? undefined
+  const transcriptionID = transcript.transcription_id ?? undefined
+  return new TranscriptionResponse({
+    text: transcript.text,
+    segments: diarizes(context.request) ? speakerTurns(tokens) : undefined,
+    words: tokens.filter(isTimedWord).map((word) => ({
+      text: word.text,
+      startSeconds: word.start,
+      endSeconds: word.end,
+      speaker: word.speaker_id ?? undefined,
+      confidence: typeof word.logprob === "number" ? Math.exp(word.logprob) : undefined,
+    })),
+    language: transcript.language_code?.toLowerCase(),
+    durationSeconds: duration,
+    usage: duration === undefined ? undefined : { type: "seconds", seconds: duration },
+    providerMetadata: transcriptionID === undefined ? undefined : { elevenlabs: { transcriptionId: transcriptionID } },
+  })
+})
+
+// ---------------------------------------------------------------------------
+// 7. Protocol and route
+// ---------------------------------------------------------------------------
+
+export const protocol = MediaProtocol.inline<Request, TranscriptionResponse>(route, {
+  unsupported: ["prompt"],
+  body: { from: fromRequest },
+  response: { decode: decodeResponse },
+})
+
+export const model = (input: MediaRoute.ModelInput) =>
+  TranscriptionModel.fromRoute<ElevenLabsTranscriptionOptions>(
+    { protocol, baseURL: DEFAULT_BASE_URL, path: PATH },
+    input,
+  )
+
+export const ElevenLabsTranscription = {
+  protocol,
+  model,
+} as const
@@ -49,7 +49,7 @@ const QueueResult = Schema.StructWithRest(
 // ---------------------------------------------------------------------------

 const sizing = (model: string) => {
-  if (/^fal-ai\/(nano-banana|flux-pro\/v1\.1-ultra)/.test(model)) return "aspect_ratio"
+  if (/^fal-ai\/(nano-banana|flux-pro\/(v1\.1-ultra|kontext))/.test(model)) return "aspect_ratio"
  if (model.startsWith("fal-ai/flux")) return "image_size"
  return undefined
 }
@@ -63,20 +63,24 @@ const validate = (request: Request) => {
    return Effect.fail(route.unsupported("media.size", `${id} sizes by aspectRatio`))
  if (request.aspectRatio !== undefined && field === "image_size")
    return Effect.fail(route.unsupported("media.aspectRatio", `${id} sizes by size (image_size)`))
-  if ((request.images?.length ?? 0) > 1 && !isEdit(id))
+  if ((request.images?.length ?? 0) > 1 && !takesImageList(id))
    return Effect.fail(
-      route.unsupported("media.images", `${id} takes one image_url; use an /edit endpoint for several images`),
+      route.unsupported(
+        "media.images",
+        `${id} takes one image_url; use an /edit or /multi endpoint for several images`,
+      ),
    )
  return Effect.void
 }

-// `/edit` endpoints take an `image_urls` list; image-to-image, fill, and Ultra take one `image_url` (beside `mask_url`).
-const isEdit = (model: string) => model.endsWith("/edit")
+// `/edit` and `/multi` (Kontext) endpoints take an `image_urls` list; image-to-image, fill, and Ultra take one
+// `image_url` (beside `mask_url`).
+const takesImageList = (model: string) => model.endsWith("/edit") || model.endsWith("/multi")

 const fromRequest = Effect.fn("FalImages.fromRequest")(function* (request: Request) {
  yield* validate(request)
  const images = yield* Effect.forEach(request.images ?? [], (image) => FalQueue.mediaUrl(image, route.name))
-  const edit = isEdit(request.model.id)
+  const list = takesImageList(request.model.id)
  return MediaProtocol.json(
    mergeJsonRecords(
      {
@@ -86,8 +90,8 @@ const fromRequest = Effect.fn("FalImages.fromRequest")(function* (request: Reque
        image_size: request.size === undefined ? undefined : MediaInput.dimensions(request.size),
        aspect_ratio: request.aspectRatio,
        output_format: request.format,
-        image_urls: edit && images.length > 0 ? images : undefined,
-        image_url: edit ? undefined : images[0],
+        image_urls: list && images.length > 0 ? images : undefined,
+        image_url: list ? undefined : images[0],
        mask_url: request.mask === undefined ? undefined : yield* FalQueue.mediaUrl(request.mask, route.name),
      },
      request.providerOptions,
@@ -112,12 +116,14 @@ const decodeResult = Effect.fn("FalImages.decodeResult")(function* (
  // With the safety checker on, flagged images come back blacked out rather than omitted.
  const flagged = (has_nsfw_concepts ?? []).flatMap((value, index) => (value ? [index] : []))
  return new ImageResponse({
-    images: images.map((image) =>
-      Media.url(image.url, {
-        mediaType: image.content_type ?? undefined,
-        info: { width: image.width ?? undefined, height: image.height ?? undefined },
-      }),
-    ),
+    images: images.map((image) => {
+      const info = { width: image.width ?? undefined, height: image.height ?? undefined }
+      // `sync_mode: true` returns data URIs instead of hosted URLs.
+      return (
+        Media.parseDataUrl(image.url, { info }) ??
+        Media.url(image.url, { mediaType: image.content_type ?? undefined, info })
+      )
+    }),
    notices:
      flagged.length === 0
        ? undefined
@@ -526,19 +526,7 @@ const mapFinishReason = (finishReason: string | undefined, hasToolCalls: boolean
  if (finishReason === undefined) return hasToolCalls ? "tool-calls" : "unknown"
  if (finishReason === "STOP") return hasToolCalls ? "tool-calls" : "stop"
  if (finishReason === "MAX_TOKENS") return "length"
-  if (
-    finishReason === "IMAGE_SAFETY" ||
-    finishReason === "RECITATION" ||
-    finishReason === "SAFETY" ||
-    finishReason === "BLOCKLIST" ||
-    finishReason === "PROHIBITED_CONTENT" ||
-    finishReason === "SPII" ||
-    finishReason === "MODEL_ARMOR" ||
-    finishReason === "IMAGE_PROHIBITED_CONTENT" ||
-    finishReason === "IMAGE_RECITATION" ||
-    finishReason === "LANGUAGE"
-  )
-    return "content-filter"
+  if (GeminiGenerateContent.contentFiltered(finishReason)) return "content-filter"
  if (
    finishReason === "MALFORMED_FUNCTION_CALL" ||
    finishReason === "UNEXPECTED_TOOL_CALL" ||
@@ -101,7 +101,7 @@ const generationConfig = (request: Request) => {
 const fromRequest = Effect.fn("GoogleImages.fromRequest")(function* (request: Request) {
  if (request.n !== undefined && request.n > 1)
    return yield* route.unsupported(
-      "image.n",
+      "media.n",
      `${route.name} generates one image per request; call it once per image instead of n=${request.n}`,
    )
  const parts = yield* Effect.forEach(request.images ?? [], (image) =>
@@ -56,10 +56,18 @@ interface State extends SpeechStream.Audio, GeminiGenerateContent.Metadata {
 // ---------------------------------------------------------------------------

 const fromRequest = Effect.fn("GoogleSpeech.fromRequest")(function* (request: MediaProtocol.Addressed<Request>) {
+  // Not in `unsupported`: that list would also reject `timestamps: false`, which asks for nothing.
+  if (request.timestamps === true)
+    return yield* route.unsupported("media.timestamps", `${route.name} does not return timestamps`)
+  if (request.format === "pcm" && request.mode === "generate" && /^gemini-3\.8-.*-tts(?:-|$)/.test(request.model.id))
+    return yield* route.unsupported(
+      "media.format",
+      `${route.name} returns WAV by default for Gemini 3.8 TTS unary requests; omit the format to accept it`,
+    )
  if (request.format !== undefined && request.format !== "pcm")
    return yield* route.unsupported(
      "media.format",
-      `${route.name} only returns raw PCM; request format "pcm" or omit it, then wrap the samples yourself`,
+      `${route.name} only accepts raw PCM as an explicit format; omit it to accept the provider's default output`,
    )
  const voiceName = SpeechStream.voiceID(request.voice)
  return MediaProtocol.json(
@@ -94,16 +102,29 @@ const step = Effect.fn("GoogleSpeech.step")(function* (state: State, frame: stri
    part.inlineData === undefined ? [] : [part.inlineData],
  )
  const next: State = { ...GeminiGenerateContent.track(state, chunk), mimeType: state.mimeType ?? audio[0]?.mimeType }
-  return [next, audio.flatMap((part) => SpeechStream.delta(next, part.data)[1])] as const
+  const events = audio.flatMap((part) => SpeechStream.delta(next, part.data)[1])
+  const withheld = next.chunks.length === 0 ? GeminiGenerateContent.withheld(route.name, chunk, frame) : undefined
+  if (withheld !== undefined) return yield* withheld
+  return [next, events] as const
 })

-const finish = (state: State) => {
+const finish = (state: State, context: MediaProtocol.ResponseContext<Request>) => {
+  if (state.finishReason === undefined) return Effect.fail(route.incomplete())
  const sampleRate = SpeechStream.sampleRate(state.mimeType) ?? DEFAULT_SAMPLE_RATE
+  const output =
+    state.mimeType?.split(";")[0]?.toLowerCase() === "audio/wav"
+      ? SpeechStream.container("wav", sampleRate)
+      : SpeechStream.pcm("pcm_s16le", sampleRate, state.mimeType ?? `audio/L16;codec=pcm;rate=${sampleRate}`)
+  if (context.request.format === "pcm" && output.info.format !== "pcm")
+    return Effect.fail(
+      route.frameError(`Google Speech returned ${output.info.format} instead of the requested raw PCM`),
+    )
  return SpeechStream.finish(route, state, {
-    ...SpeechStream.pcm("pcm_s16le", sampleRate, state.mimeType ?? `audio/L16;codec=pcm;rate=${sampleRate}`),
+    ...output,
    usage: GeminiGenerateContent.usage(state.usage),
+    notices: GeminiGenerateContent.notices(route.name, state),
    providerMetadata: GeminiGenerateContent.providerMetadata(state),
-    detail: state.finishReason === undefined ? undefined : `finish reason: ${state.finishReason}`,
+    detail: `finish reason: ${state.finishReason}`,
  })
 }

@@ -112,7 +133,7 @@ const finish = (state: State) => {
 // ---------------------------------------------------------------------------

 export const protocol = MediaProtocol.stream<Request, SpeechEvent, string, State>(route, {
-  unsupported: ["instructions", "speed", "timestamps"],
+  unsupported: ["instructions", "speed"],
  body: { from: fromRequest },
  frames: (bytes, context) => GeminiGenerateContent.frames(bytes, context.request.mode),
  initial: () => ({ chunks: [] }),
@@ -154,6 +154,9 @@ const step = Effect.fn("GoogleTranscription.step")(function* (state: State, fram
    .filter((item) => item.length > 0)
    .join(" ")
  const delta = text.length === 0 || state.text.length === 0 ? text : ` ${text}`
+  const withheld =
+    state.text.length + delta.length === 0 ? GeminiGenerateContent.withheld(route.name, chunk, frame) : undefined
+  if (withheld !== undefined) return yield* withheld
  const events: ReadonlyArray<TranscriptionEvent> = [
    ...(delta.length === 0 ? [] : [TranscriptionTextDeltaEvent.make({ delta })]),
    ...segments.map((segment) => TranscriptionSegmentEvent.make({ segment })),
@@ -169,6 +172,7 @@ const finish = (state: State) => {
      segments: state.segments.length === 0 ? undefined : state.segments,
      words: state.words.length === 0 ? undefined : state.words,
      usage: GeminiGenerateContent.usage(state.usage),
+      notices: GeminiGenerateContent.notices(route.name, state),
      providerMetadata: GeminiGenerateContent.providerMetadata(state),
    }),
  ])
@@ -36,7 +36,9 @@ const StartResponse = Schema.Struct({ name: Schema.String })

 const Operation = Schema.Struct({
  done: Schema.optional(Schema.Boolean),
-  error: Schema.optional(Schema.Struct({ message: Schema.optional(Schema.String) })),
+  error: Schema.optional(
+    Schema.Struct({ code: Schema.optional(Schema.Number), message: Schema.optional(Schema.String) }),
+  ),
  response: Schema.optional(
    Schema.Struct({
      generateVideoResponse: Schema.optional(
@@ -60,6 +62,16 @@ const Operation = Schema.Struct({
  metadata: Schema.optional(Schema.Unknown),
 })

+// Operation errors are `google.rpc.Status`; unlisted codes (INTERNAL, UNAVAILABLE, ...) are provider-side.
+const FAILURE = {
+  3: "InvalidRequest", // INVALID_ARGUMENT
+  7: "Authentication", // PERMISSION_DENIED
+  8: "RateLimit", // RESOURCE_EXHAUSTED
+  9: "InvalidRequest", // FAILED_PRECONDITION
+  11: "InvalidRequest", // OUT_OF_RANGE
+  16: "Authentication", // UNAUTHENTICATED
+} as const satisfies Record<number, MediaProtocol.Failure>
+
 // ---------------------------------------------------------------------------
 // 5. Request body construction
 // ---------------------------------------------------------------------------
@@ -149,12 +161,12 @@ const decodeResult = Effect.fn("GoogleVideo.decodeResult")(function* (
  const output = yield* decodeOperation(response)
  const operation = output.value
  const status = statusOf(operation)
-  if (status === "running")
-    return yield* output.invalid(`${route.name} operation ${context.token.operation} has not finished`)
+  if (status === "running") return yield* output.pending(context.token.operation)
  if (status === "failed")
    return yield* output.ended(
      "failed",
      `${route.name} operation failed${operation.error?.message === undefined ? "" : `: ${operation.error.message}`}`,
+      MediaProtocol.failure(FAILURE, operation.error?.code),
    )
  const generated = operation.response?.generateVideoResponse
  // Downloads require the same API key as the poll; the asset carries it transiently and follows the redirect.
@@ -1,5 +1,6 @@
 export * as AnthropicMessages from "./anthropic-messages.js"
 export * as BedrockConverse from "./bedrock-converse.js"
+export * as CohereChat from "./cohere-chat.js"
 export * as Gemini from "./gemini.js"
 export * as MistralChat from "./mistral-chat.js"
 export * as OpenAIChat from "./openai-chat.js"
@@ -10,14 +10,19 @@ const WebSearch = Schema.Struct({
  name: Schema.Literal("web_search"),
  user_location: MetaResponses.WebSearch.fields.user_location,
 })
+const MetaCacheControl = Schema.Struct({
+  type: Schema.tag("ephemeral"),
+  ttl: Schema.optional(Schema.Literals(["5m", "1h"])),
+})
+const FunctionTool = Schema.Struct({
+  name: Schema.String,
+  description: Schema.String,
+  input_schema: JsonObject,
+  cache_control: Schema.optional(MetaCacheControl),
+})
 const Body = Schema.Struct({
  ...AnthropicMessages.AnthropicMessagesBody.fields,
-  tools: optionalArray(
-    Schema.Union([
-      Schema.Struct({ name: Schema.String, description: Schema.String, input_schema: JsonObject }),
-      WebSearch,
-    ]),
-  ),
+  tools: optionalArray(Schema.Union([FunctionTool, WebSearch])),
 })

 const fromRequest = Effect.fn("MetaMessages.fromRequest")(function* (request: LLMRequest) {
@@ -426,6 +426,7 @@ interface ActiveContent {
  readonly type: "text" | "reasoning"
  readonly id: string
  readonly thinking?: MistralThinkingContent
+  readonly thinkingUnits?: MistralThinkingUnit[]
 }

 export interface ParserState {
@@ -502,8 +503,8 @@ const closeActive = (state: ParserState, events: LLMEvent[]) => {
          state.lifecycle,
          events,
          state.active.id,
-          thinkingMetadata(state.active.thinking ?? { type: "thinking", thinking: [] }),
-          thinkingText(state.active.thinking?.thinking ?? []),
+          thinkingMetadata({ ...state.active.thinking, type: "thinking", thinking: state.active.thinkingUnits ?? [] }),
+          thinkingText(state.active.thinkingUnits ?? []),
        )
  return { ...state, lifecycle, active: undefined }
 }
@@ -524,20 +525,23 @@ const appendThinking = (state: ParserState, events: LLMEvent[], part: MistralOut
  const current = state.active?.type === "reasoning" ? state : closeActive(state, events)
  const units = thinkingUnits(part.thinking)
  const active = current.active ?? { type: "reasoning" as const, id: `reasoning-${current.nextContent}` }
+  // Keep native units out of streamed events until the block is complete.
+  const accumulated = active.thinkingUnits ?? []
+  accumulated.push(...units)
  const thinking = {
    ...active.thinking,
    ...part,
    type: "thinking" as const,
-    thinking: [...(active.thinking?.thinking ?? []), ...units],
+    thinking: [],
  }
  const text = thinkingText(units)
  return {
    ...current,
    lifecycle:
      text.length > 0
-        ? Lifecycle.reasoningDelta(current.lifecycle, events, active.id, text, thinkingMetadata(thinking))
-        : Lifecycle.reasoningStart(current.lifecycle, events, active.id, thinkingMetadata(thinking)),
-    active: { ...active, thinking },
+        ? Lifecycle.reasoningDelta(current.lifecycle, events, active.id, text)
+        : Lifecycle.reasoningStart(current.lifecycle, events, active.id),
+    active: { ...active, thinking, thinkingUnits: accumulated },
    nextContent: current.active ? current.nextContent : current.nextContent + 1,
  }
 }
@@ -20,7 +20,6 @@ import {
  type LLMRequest,
  type MediaPart,
  type ReasoningPart,
-  type TextPart,
  type ToolCallPart,
  type ToolDefinition,
 } from "../schema/index.js"
@@ -45,6 +44,7 @@ const OpenAIChatCacheControl = Schema.Struct({
  type: Schema.Literal("ephemeral"),
  ttl: Schema.optional(Schema.String),
 })
+type OpenAIChatCacheControl = Schema.Schema.Type<typeof OpenAIChatCacheControl>

 const OpenAIChatFunction = Schema.Struct({
  name: Schema.String,
@@ -64,6 +64,14 @@ const OpenAIChatTool = Schema.Struct({
 })
 type OpenAIChatTool = Schema.Schema.Type<typeof OpenAIChatTool>

+// Gemini's OpenAI-compatible surface carries thought signatures in tool call
+// `extra_content` and rejects replayed parallel calls without them:
+// https://ai.google.dev/gemini-api/docs/thinking#signatures
+const ExtraContent = Schema.Struct({
+  google: Schema.Struct({ thought_signature: Schema.String }),
+})
+const decodeExtraContent = (value: unknown) => Option.getOrUndefined(Schema.decodeUnknownOption(ExtraContent)(value))
+
 const OpenAIChatAssistantToolCall = Schema.Struct({
  id: Schema.String,
  type: Schema.tag("function"),
@@ -71,6 +79,7 @@ const OpenAIChatAssistantToolCall = Schema.Struct({
    name: Schema.String,
    arguments: Schema.String,
  }),
+  extra_content: Schema.optional(ExtraContent),
 })
 type OpenAIChatAssistantToolCall = Schema.Schema.Type<typeof OpenAIChatAssistantToolCall>

@@ -112,18 +121,15 @@ const decodeReasoningDetail = Schema.decodeUnknownOption(ReasoningDetail)
 const knownReasoningDetails = (details: ReadonlyArray<unknown>) =>
  details.flatMap((detail) => Option.toArray(decodeReasoningDetail(detail)))

-// Intentionally omit Gemini's provider-specific `extra_content.google.thought_signature`
-// extension until direct Google OpenAI-compatible routing is supported here:
-// https://github.com/vercel/ai/issues/11590
-// https://github.com/vercel/ai/pull/11745
-// https://ai.google.dev/gemini-api/docs/thought-signatures#openai
+const OpenAIChatTextContent = Schema.Struct({
+  type: Schema.Literal("text"),
+  text: Schema.String,
+  cache_control: Schema.optional(OpenAIChatCacheControl),
+})
+type OpenAIChatTextContent = Schema.Schema.Type<typeof OpenAIChatTextContent>

 const OpenAIChatUserContent = Schema.Union([
-  Schema.Struct({
-    type: Schema.Literal("text"),
-    text: Schema.String,
-    cache_control: Schema.optional(OpenAIChatCacheControl),
-  }),
+  OpenAIChatTextContent,
  Schema.Struct({
    type: Schema.Literal("image_url"),
    image_url: Schema.Struct({ url: Schema.String }),
@@ -133,6 +139,7 @@ const OpenAIChatUserContent = Schema.Union([
    file: Schema.Struct({ filename: Schema.String, file_data: Schema.String }),
  }),
 ])
+type OpenAIChatUserContent = Schema.Schema.Type<typeof OpenAIChatUserContent>

 const OpenAIChatMessage = Schema.Union([
  Schema.Struct({
@@ -146,21 +153,19 @@ const OpenAIChatMessage = Schema.Union([
  Schema.StructWithRest(
    Schema.Struct({
      role: Schema.Literal("assistant"),
-      content: Schema.NullOr(Schema.String),
+      content: Schema.NullOr(Schema.Union([Schema.String, Schema.Array(OpenAIChatTextContent)])),
      tool_calls: optionalArray(OpenAIChatAssistantToolCall),
      reasoning_content: Schema.optional(Schema.String),
      reasoning: Schema.optional(Schema.String),
      reasoning_text: Schema.optional(Schema.String),
      reasoning_details: Schema.optional(Schema.Unknown),
-      cache_control: Schema.optional(OpenAIChatCacheControl),
    }),
    [Schema.Record(Schema.String, Schema.Unknown)],
  ),
  Schema.Struct({
    role: Schema.Literal("tool"),
    tool_call_id: Schema.String,
-    content: Schema.String,
-    cache_control: Schema.optional(OpenAIChatCacheControl),
+    content: Schema.Union([Schema.String, Schema.Array(OpenAIChatTextContent)]),
  }),
 ]).pipe(Schema.toTaggedUnion("role"))
 type OpenAIChatMessage = Schema.Schema.Type<typeof OpenAIChatMessage>
@@ -207,9 +212,11 @@ const OpenAIChatUsage = Schema.StructWithRest(
    prompt_tokens: optionalNull(Schema.Number),
    completion_tokens: optionalNull(Schema.Number),
    total_tokens: optionalNull(Schema.Number),
-    // Zai reports cache hits as top-level `cached_tokens`; DeepSeek uses `prompt_cache_hit_tokens`.
+    // Provider-specific cache accounting fields.
    cached_tokens: optionalNull(Schema.Number),
    prompt_cache_hit_tokens: optionalNull(Schema.Number),
+    cache_read_input_tokens: optionalNull(Schema.Number),
+    cache_created_input_tokens: optionalNull(Schema.Number),
    prompt_tokens_details: optionalNull(
      Schema.StructWithRest(
        Schema.Struct({
@@ -242,6 +249,7 @@ const OpenAIChatToolCallDelta = Schema.Struct({
  index: optionalNull(Schema.Number),
  id: optionalNull(Schema.String),
  function: optionalNull(OpenAIChatToolCallDeltaFunction),
+  extra_content: optionalNull(Schema.Unknown),
 })
 type OpenAIChatToolCallDelta = Schema.Schema.Type<typeof OpenAIChatToolCallDelta>

@@ -294,6 +302,7 @@ interface PendingToolDelta {
  readonly id?: string
  readonly name?: string
  readonly input: string
+  readonly extraContent?: Schema.Schema.Type<typeof ExtraContent>
 }

 export interface ParserState {
@@ -322,9 +331,7 @@ export interface ParserState {
 // OpenAI Chat wire format. Keep provider quirks here instead of leaking native
 // fields into `LLMRequest`.
 interface LoweringOptions {
-  readonly cacheControl?: (
-    cache: CacheHint | undefined,
-  ) => Schema.Schema.Type<typeof OpenAIChatCacheControl> | undefined
+  readonly cacheControl?: (cache: CacheHint | undefined) => OpenAIChatCacheControl | undefined
  readonly toolCallID?: (id: string) => string
 }

@@ -347,13 +354,17 @@ const lowerToolChoice = (toolChoice: NonNullable<LLMRequest["toolChoice"]>) =>
    tool: (name) => ({ type: "function" as const, function: { name } }),
  })

-const lowerToolCall = (part: ToolCallPart, options: LoweringOptions): OpenAIChatAssistantToolCall => ({
+const lowerToolCall = (
+  part: ToolCallPart,
+  options: LoweringOptions & { readonly providerMetadataKey: string },
+): OpenAIChatAssistantToolCall => ({
  id: options.toolCallID?.(part.id) ?? part.id,
  type: "function",
  function: {
    name: part.name,
    arguments: ProviderShared.encodeJson(part.input === undefined ? {} : part.input),
  },
+  extra_content: decodeExtraContent(part.providerMetadata?.[options.providerMetadataKey]?.extraContent),
 })

 const lowerMedia = Effect.fn("OpenAIChat.lowerMedia")(function* (part: MediaPart) {
@@ -406,7 +417,7 @@ const lowerUserMessage = Effect.fn("OpenAIChat.lowerUserMessage")(function* (
  message: OpenAIChatRequestMessage,
  options: LoweringOptions,
 ) {
-  const content: Array<Schema.Schema.Type<typeof OpenAIChatUserContent>> = []
+  const content: OpenAIChatUserContent[] = []
  for (const part of message.content) {
    if (part.type === "text") {
      content.push({ type: "text", text: part.text, cache_control: options.cacheControl?.(part.cache) })
@@ -432,14 +443,14 @@ const lowerAssistantMessage = Effect.fn("OpenAIChat.lowerAssistantMessage")(func
  requireReasoning: boolean,
  options: LoweringOptions & { readonly providerMetadataKey: string },
 ) {
-  const content: TextPart[] = []
+  const content: OpenAIChatTextContent[] = []
  const reasoning: ReasoningPart[] = []
  const toolCalls: OpenAIChatAssistantToolCall[] = []
  for (const part of message.content) {
    if (!ProviderShared.supportsContent(part, ["text", "reasoning", "tool-call"]))
      return yield* ProviderShared.unsupportedContent("OpenAI Chat", "assistant", ["text", "reasoning", "tool-call"])
    if (part.type === "text") {
-      content.push(part)
+      content.push({ type: "text", text: part.text, cache_control: options.cacheControl?.(part.cache) })
      continue
    }
    if (part.type === "reasoning") {
@@ -477,14 +488,15 @@ const lowerAssistantMessage = Effect.fn("OpenAIChat.lowerAssistantMessage")(func
    if (reasoning.length === 0) return nativeReasoning ?? (requireReasoning ? "" : undefined)
    return text
  })()
-  const cached = message.content.findLast((part) => "cache" in part && part.cache !== undefined)
-  const cacheControl = options.cacheControl?.(cached && "cache" in cached ? cached.cache : undefined)
  const result = {
    role: "assistant" as const,
-    content: content.length > 0 ? content.map((part) => part.text).join("") : toolCalls.length > 0 ? null : "",
+    content: (() => {
+      if (content.some((part) => part.cache_control !== undefined)) return content
+      if (content.length === 0 && toolCalls.length > 0) return null
+      return content.map((part) => part.text).join("")
+    })(),
    ...(toolCalls.length > 0 ? { tool_calls: toolCalls } : {}),
    ...(details !== undefined ? { reasoning_details: details } : {}),
-    ...(cacheControl !== undefined ? { cache_control: cacheControl } : {}),
  }
  if (field === undefined || reasoningText === undefined) return result
  return { ...result, [field]: reasoningText }
@@ -495,33 +507,38 @@ const lowerToolMessages = Effect.fn("OpenAIChat.lowerToolMessages")(function* (
  options: LoweringOptions,
 ) {
  const messages: OpenAIChatMessage[] = []
-  const attachments: Array<Schema.Schema.Type<typeof OpenAIChatUserContent>> = []
+  const attachments: OpenAIChatUserContent[] = []
  for (const part of message.content) {
    if (!ProviderShared.supportsContent(part, ["tool-result"]))
      return yield* ProviderShared.unsupportedContent("OpenAI Chat", "tool", ["tool-result"])
    if (part.result.type !== "content") {
-      messages.push({
-        role: "tool",
-        tool_call_id: options.toolCallID?.(part.id) ?? part.id,
-        content: ProviderShared.toolResultText(part),
-        cache_control: options.cacheControl?.(part.cache),
-      })
+      messages.push(
+        toolMessage(
+          options.toolCallID?.(part.id) ?? part.id,
+          ProviderShared.toolResultText(part),
+          options.cacheControl?.(part.cache),
+        ),
+      )
      continue
    }
    const content: ReadonlyArray<Tool.Content> = part.result.value
    const text = content.filter((item) => item.type === "text").map((item) => item.text)
-    messages.push({
-      role: "tool",
-      tool_call_id: options.toolCallID?.(part.id) ?? part.id,
-      content: text.join("\n"),
-      cache_control: options.cacheControl?.(part.cache),
-    })
+    messages.push(
+      toolMessage(options.toolCallID?.(part.id) ?? part.id, text.join("\n"), options.cacheControl?.(part.cache)),
+    )
    const files = content.filter((item) => item.type === "file")
    attachments.push(...(yield* Effect.forEach(files, (item) => lowerMedia(ProviderShared.toolFileMedia(item)))))
  }
  return { messages, attachments }
 })

+// Chat cache breakpoints belong on text content parts, not on the message itself.
+const toolMessage = (toolCallID: string, text: string, cacheControl: OpenAIChatCacheControl | undefined) => ({
+  role: "tool" as const,
+  tool_call_id: toolCallID,
+  content: cacheControl === undefined ? text : [{ type: "text" as const, text, cache_control: cacheControl }],
+})
+
 const lowerMessage = Effect.fn("OpenAIChat.lowerMessage")(function* (
  message: OpenAIChatRequestMessage,
  reasoningField: string | undefined,
@@ -580,7 +597,7 @@ const lowerMessages = Effect.fn("OpenAIChat.lowerMessages")(function* (request:
    if (requireAssistantAfterTool && messages.at(-1)?.role === "tool")
      messages.push({ role: "assistant", content: "Done." })
  }
-  const pendingAttachments: Array<Schema.Schema.Type<typeof OpenAIChatUserContent>> = []
+  const pendingAttachments: OpenAIChatUserContent[] = []
  const flushAttachments = () => {
    if (pendingAttachments.length === 0) return
    bridgeTools()
@@ -590,24 +607,25 @@ const lowerMessages = Effect.fn("OpenAIChat.lowerMessages")(function* (request:
    if (message.role === "user") bridgeTools()
    if (message.role === "system") {
      const part = yield* ProviderShared.wrappedSystemUpdate("OpenAI Chat", message)
+      const cacheControl = options.cacheControl?.(part.cache)
      if (pendingAttachments.length > 0) {
        messages.push({
          role: "user",
          content: [
            ...pendingAttachments.splice(0),
-            { type: "text", text: part.text, cache_control: options.cacheControl?.(part.cache) },
+            { type: "text", text: part.text, cache_control: cacheControl },
          ],
        })
        continue
      }
      const previous = messages.at(-1)
      if (previous?.role === "user" && typeof previous.content === "string")
-        messages[messages.length - 1] = options.cacheControl?.(part.cache)
+        messages[messages.length - 1] = cacheControl
          ? {
              role: "user",
              content: [
                { type: "text", text: previous.content },
-                { type: "text", text: part.text, cache_control: options.cacheControl(part.cache) },
+                { type: "text", text: part.text, cache_control: cacheControl },
              ],
            }
          : { role: "user", content: `${previous.content}\n${part.text}` }
@@ -616,15 +634,15 @@ const lowerMessages = Effect.fn("OpenAIChat.lowerMessages")(function* (request:
          role: "user",
          content: [
            ...previous.content,
-            { type: "text", text: part.text, cache_control: options.cacheControl?.(part.cache) },
+            { type: "text", text: part.text, cache_control: cacheControl },
          ],
        }
      else
        messages.push(
-          options.cacheControl?.(part.cache)
+          cacheControl
            ? {
                role: "user",
-                content: [{ type: "text", text: part.text, cache_control: options.cacheControl(part.cache) }],
+                content: [{ type: "text", text: part.text, cache_control: cacheControl }],
              }
            : { role: "user", content: part.text },
        )
@@ -721,7 +739,9 @@ const detectSupportsStore = (provider: string, baseURL: string | undefined): boo
    p === "vercel-ai-gateway" || url.includes("ai-gateway.vercel.sh") || url.includes("vercel.sh")
  const isAntLing = p === "ant-ling" || url.includes("api.ant-ling.com")
  const isOpencode = p === "opencode" || url.includes("opencode.ai")
+  const isGemini = url.includes("generativelanguage.googleapis.com")
  const isNonStandard =
+    isGemini ||
    isNvidia ||
    isCerebras ||
    isXai ||
@@ -885,16 +905,19 @@ const mapFinishReason = Effect.fn("OpenAIChat.mapFinishReason")(function* (event
 // satisfied on both sides.
 // Providers differ on cache-hit location: OpenAI uses
 // `prompt_tokens_details.cached_tokens`, DeepSeek uses
-// `prompt_cache_hit_tokens`, and Zai uses top-level `cached_tokens`.
+// `prompt_cache_hit_tokens`, Zai uses top-level `cached_tokens`, and
+// DigitalOcean uses top-level `cache_read_input_tokens` / `cache_created_input_tokens`.
 const mapUsage = (usage: OpenAIChatEvent["usage"], providerMetadataKey: string): Usage | undefined => {
  if (!usage) return undefined
  const input = usage.prompt_tokens ?? undefined
  const output = usage.completion_tokens ?? undefined
-  const cached = (usage.prompt_tokens_details?.cached_tokens ??
-    (usage as { prompt_cache_hit_tokens?: number | null }).prompt_cache_hit_tokens ??
-    (usage as { cached_tokens?: number | null }).cached_tokens ??
-    undefined) as number | undefined
-  const cacheWrite = usage.prompt_tokens_details?.cache_write_tokens ?? undefined
+  const cached =
+    usage.prompt_tokens_details?.cached_tokens ??
+    usage.prompt_cache_hit_tokens ??
+    usage.cached_tokens ??
+    usage.cache_read_input_tokens ??
+    undefined
+  const cacheWrite = usage.prompt_tokens_details?.cache_write_tokens ?? usage.cache_created_input_tokens ?? undefined
  const reasoning = usage.completion_tokens_details?.reasoning_tokens ?? undefined
  const nonCached = ProviderShared.subtractTokens(input, ProviderShared.sumTokens(cached, cacheWrite))
  return new Usage({
@@ -1114,12 +1137,13 @@ const step = (state: ParserState, event: OpenAIChatEvent) =>
      const id = current?.id ?? pending?.id ?? (tool.id || undefined)
      const name = current?.name ?? pending?.name ?? (tool.function?.name || undefined)
      const text = `${pending?.input ?? ""}${tool.function?.arguments ?? ""}`
+      const extraContent = pending?.extraContent ?? decodeExtraContent(tool.extra_content)
      latestToolIndex = index
      nextToolIndex = Math.max(nextToolIndex, index + 1)
      if (!current && (!id || !name)) {
        pendingTools = {
          ...pendingTools,
-          [index]: { id: id || undefined, name: name || undefined, input: text },
+          [index]: { id: id || undefined, name: name || undefined, input: text, extraContent },
        }
        continue
      }
@@ -1131,7 +1155,12 @@ const step = (state: ParserState, event: OpenAIChatEvent) =>
        ADAPTER,
        tools,
        index,
-        { id: id || undefined, name: name || undefined, text },
+        {
+          id: id || undefined,
+          name: name || undefined,
+          text,
+          providerMetadata: extraContent && { [state.providerMetadataKey]: { extraContent } },
+        },
        "OpenAI Chat tool call delta is missing id or name",
      )
      if (ToolStream.isError(result))
@@ -48,15 +48,17 @@ const Usage = Schema.Struct({
  output_tokens_details: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)),
 })

-const OpenAIImageResponse = Schema.Struct({
-  data: Schema.Array(
-    Schema.Struct({
-      b64_json: Schema.optional(Schema.String),
-      url: Schema.optional(Schema.String),
-      revised_prompt: Schema.optional(Schema.String),
-    }),
-  ),
+/** What the provider actually rendered; it can differ from the request when `auto` or a default applied. */
+const Settings = {
  output_format: Schema.optional(Schema.String),
+  size: Schema.optional(Schema.String),
+  quality: Schema.optional(Schema.String),
+  background: Schema.optional(Schema.String),
+}
+
+const OpenAIImageResponse = Schema.Struct({
+  data: Schema.Array(Schema.Struct({ b64_json: Schema.String })),
+  ...Settings,
  usage: Schema.optional(Usage),
 })

@@ -69,11 +71,13 @@ const StreamEvent = Schema.Union([
    type: Schema.Literals(["image_generation.partial_image", "image_edit.partial_image"]),
    b64_json: Schema.String,
    partial_image_index: Schema.Number,
+    ...Settings,
    output_format: Schema.String,
  }),
  Schema.Struct({
    type: Schema.Literals(["image_generation.completed", "image_edit.completed"]),
    b64_json: Schema.String,
+    ...Settings,
    output_format: Schema.String,
    usage: Schema.optional(Usage),
  }),
@@ -92,6 +96,9 @@ type Frame = string | { readonly document: string; readonly requested: string |
 interface State {
  readonly completed: number
  readonly format?: string
+  readonly size?: string
+  readonly quality?: string
+  readonly background?: string
  readonly usage?: MediaUsage
 }

@@ -110,10 +117,6 @@ const nativeOptions = (options: OpenAIImageOptions | undefined) => {

 const streamOptions = (request: MediaProtocol.Addressed<Request>) => {
  if (request.mode !== "stream") return Effect.succeed(undefined)
-  if (request.model.id.startsWith("dall-e"))
-    return Effect.fail(
-      route.unsupported("media.stream", `${request.model.id} does not stream; use Image.generate or a GPT image model`),
-    )
  if (request.n !== undefined && request.n > 1)
    return Effect.fail(
      route.unsupported("media.n", `${route.name} streams one image; use Image.generate for n=${request.n}`),
@@ -194,21 +197,34 @@ const usage = (value: Schema.Schema.Type<typeof Usage> | undefined): MediaUsage
        details: { openai: value },
      }

-const eventImage = (frame: string, label: string, data: string, format: string) =>
+/** `size` echoes the rendered `WIDTHxHEIGHT`; `auto` or any other value leaves the dimensions unknown. */
+const info = (format: string, size: string | undefined): Media.Info => {
+  const match = size?.match(/^(\d+)x(\d+)$/)
+  return match ? { format, width: Number(match[1]), height: Number(match[2]) } : { format }
+}
+
+const eventImage = (frame: string, label: string, data: string, format: string, size: string | undefined) =>
  MediaInput.decodedAsset((message, cause) => route.frameError(message, frame, cause), label, data, `image/${format}`, {
-    info: { format },
+    info: info(format, size),
  })

 const onEvent = Effect.fn("OpenAIImages.onEvent")(function* (state: State, frame: string) {
  const event = yield* decodeEvent(frame)
  const format = event.output_format
  if ("partial_image_index" in event) {
-    const image = yield* eventImage(frame, `${route.name} partial image`, event.b64_json, format)
+    const image = yield* eventImage(frame, `${route.name} partial image`, event.b64_json, format, event.size)
    return [state, [ImagePartialEvent.make({ index: event.partial_image_index, image })]] as const
  }
-  const image = yield* eventImage(frame, `${route.name} result ${state.completed}`, event.b64_json, format)
+  const image = yield* eventImage(frame, `${route.name} result ${state.completed}`, event.b64_json, format, event.size)
  return [
-    { ...state, completed: state.completed + 1, format, usage: usage(event.usage) },
+    {
+      completed: state.completed + 1,
+      format,
+      size: event.size,
+      quality: event.quality,
+      background: event.background,
+      usage: usage(event.usage),
+    },
    [ImageOutputEvent.make({ index: state.completed, image })],
  ] as const
 })
@@ -219,16 +235,20 @@ const onDocument = Effect.fn("OpenAIImages.onDocument")(function* (frame: Exclud
    Effect.mapError((cause) => invalid(`${route.name} returned an invalid response`, cause)),
  )
  const format = decoded.output_format ?? frame.requested ?? "png"
-  const mediaType = `image/${format}`
  const images = yield* Effect.forEach(decoded.data, (item, index) =>
-    MediaInput.imageOutput(invalid, `${route.name} result ${index}`, item, mediaType, {
-      info: { format },
-      providerMetadata:
-        item.revised_prompt === undefined ? undefined : { openai: { revisedPrompt: item.revised_prompt } },
+    MediaInput.decodedAsset(invalid, `${route.name} result ${index}`, item.b64_json, `image/${format}`, {
+      info: info(format, decoded.size),
    }),
  )
  if (images.length === 0) return yield* invalid(`${route.name} returned no images`)
-  const state: State = { completed: images.length, format, usage: usage(decoded.usage) }
+  const state: State = {
+    completed: images.length,
+    format,
+    size: decoded.size,
+    quality: decoded.quality,
+    background: decoded.background,
+    usage: usage(decoded.usage),
+  }
  return [state, images.map((image, index) => ImageOutputEvent.make({ index, image }))] as const
 })

@@ -237,7 +257,17 @@ const step = (state: State, frame: Frame) => (typeof frame === "string" ? onEven
 const finish = (state: State) => {
  if (state.completed === 0) return Effect.fail(route.incomplete())
  return Effect.succeed([
-    ImageFinishEvent.make({ usage: state.usage, providerMetadata: { openai: { outputFormat: state.format } } }),
+    ImageFinishEvent.make({
+      usage: state.usage,
+      providerMetadata: {
+        openai: {
+          outputFormat: state.format,
+          size: state.size,
+          quality: state.quality,
+          background: state.background,
+        },
+      },
+    }),
  ])
 }

@@ -143,12 +143,15 @@ const adapter = {
  restoreHostedToolItem: (item: unknown) => (Schema.is(OpenAIResponsesHostedToolItem)(item) ? item : undefined),
 } satisfies OpenResponses.ProviderAdapter

-// Only GPT-6 Astra accepts `configuration_update`, and never alongside automatic `context_management` compaction.
+// GPT-6 and later default to `configuration_update` support, except in `reasoning.mode: "pro"`
+// or alongside automatic `context_management` compaction.
 const supportsEffortUpdates = (request: LLMRequest) => {
  if (request.providerOptions?.contextManagement !== undefined) return false
+  if (Schema.is(Schema.Struct({ mode: Schema.Literal("pro") }))(request.http?.body?.reasoning)) return false
  const override = request.model.compatibility?.supportsEffortUpdates
  if (override !== undefined) return override
-  return /(?:^|\/)gpt-6-astra$/i.test(request.model.id)
+  const match = /(?:^|\/)gpt-(\d+)(?:\.\d+)?(?:-|$)/i.exec(request.model.id)
+  return match !== null && Number(match[1]) >= 6
 }

 const nativeImageToolInput = (tool: ToolDefinition) => {
@@ -60,7 +60,17 @@ interface State extends SpeechStream.Audio {
 // `sse` is not supported for `tts-1` or `tts-1-hd`; those models stream the raw audio body instead.
 const supportsSse = (model: string) => !/^tts-1(-hd)?(-|$)/.test(model)

+const FORMATS = new Set(["mp3", "opus", "aac", "flac", "wav", "pcm"])
+
 const fromRequest = Effect.fn("OpenAISpeech.fromRequest")(function* (request: MediaProtocol.Addressed<Request>) {
+  // Not in `unsupported`: that list would also reject `timestamps: false`, which asks for nothing.
+  if (request.timestamps === true)
+    return yield* route.unsupported("media.timestamps", `${route.name} does not return timestamps`)
+  if (request.format !== undefined && !FORMATS.has(request.format))
+    return yield* route.unsupported(
+      "media.format",
+      `${route.name} supports the mp3, opus, aac, flac, wav, and pcm formats, not "${request.format}"`,
+    )
  return MediaProtocol.json(
    mergeJsonRecords(
      {
@@ -109,7 +119,9 @@ const onEvent = Effect.fn("OpenAISpeech.onEvent")(function* (state: State, frame

 const finish = (state: State, context: MediaProtocol.ResponseContext<Request>) => {
  if (isSse(context.body) && !state.done) return Effect.fail(route.incomplete())
-  const format = context.request.format ?? "mp3"
+  // The sent body reflects `providerOptions` and `http.body` overrides of `format`.
+  const sent = context.body.type === "json" ? context.body.value.response_format : undefined
+  const format = typeof sent === "string" ? sent : "mp3"
  return SpeechStream.finish(route, state, {
    ...(format === "pcm" ? SpeechStream.pcm("pcm_s16le", PCM_SAMPLE_RATE) : SpeechStream.container(format)),
    usage: state.usage,
@@ -121,7 +133,7 @@ const finish = (state: State, context: MediaProtocol.ResponseContext<Request>) =
 // ---------------------------------------------------------------------------

 export const protocol = MediaProtocol.stream<Request, SpeechEvent, string | Uint8Array, State>(route, {
-  unsupported: ["language", "timestamps"],
+  unsupported: ["language"],
  body: { from: fromRequest },
  frames: (bytes, context) => (isSse(context.body) ? Framing.sse.frame(bytes) : bytes),
  initial: () => ({ chunks: [], done: false }),
@@ -1,8 +1,9 @@
 import { Effect, Schema, Stream } from "effect"
+import { classifyProviderFailure } from "../provider-error.js"
 import { Framing } from "../route/framing.js"
 import { MediaProtocol } from "../route/media-protocol.js"
 import { MediaRoute } from "../route/media.js"
-import { mergeJsonRecords, type MediaUsage } from "../schema/index.js"
+import { AIError, mergeJsonRecords, type MediaUsage } from "../schema/index.js"
 import {
  TranscriptionFinishEvent,
  TranscriptionModel,
@@ -59,6 +60,9 @@ const Usage = Schema.Union([
    input_tokens: Schema.optional(Schema.Number),
    output_tokens: Schema.optional(Schema.Number),
    total_tokens: Schema.optional(Schema.Number),
+    input_token_details: Schema.optional(
+      Schema.Struct({ audio_tokens: Schema.optional(Schema.Number), text_tokens: Schema.optional(Schema.Number) }),
+    ),
  }),
  Schema.Struct({ type: Schema.Literal("duration"), seconds: Schema.Number }),
 ])
@@ -75,14 +79,23 @@ const transcriptFields = {
  usage: Schema.optional(Usage),
 }

+/** OpenAI may add stream event types; frames outside `EVENT_TYPES` are ignored. */
+const EventType = Schema.Struct({ type: Schema.String })
 const Event = Schema.Union([
  Schema.Struct({ type: Schema.Literal("transcript.text.delta"), delta: Schema.String }),
  Schema.Struct({ type: Schema.Literal("transcript.text.segment"), ...Segment.fields }),
  Schema.Struct({ type: Schema.Literal("transcript.text.done"), ...transcriptFields }),
+  Schema.Struct({
+    type: Schema.Literal("error"),
+    message: Schema.optional(Schema.String),
+    error: Schema.optional(Schema.Struct({ message: Schema.optional(Schema.String) })),
+  }),
 ])
+const EVENT_TYPES = new Set(["transcript.text.delta", "transcript.text.segment", "transcript.text.done", "error"])
 const Transcript = Schema.Struct(transcriptFields)
 type Transcript = Schema.Schema.Type<typeof Transcript>

+const decodeEventType = route.decodeFrame(EventType)
 const decodeEvent = route.decodeFrame(Event)
 const decodeTranscript = route.decodeFrame(Transcript)

@@ -118,10 +131,12 @@ const capabilities = (model: string): Capabilities => {
  return TRANSCRIBE
 }

+/** whisper-1 ignores `stream`, so its `stream` mode sends a plain request and emits only `finish`. */
+const streamsEvents = (request: MediaProtocol.Addressed<Request>) =>
+  request.mode === "stream" && capabilities(request.model.id).stream
+
 const validate = (request: MediaProtocol.Addressed<Request>, model: Capabilities) => {
  const id = request.model.id
-  if (request.mode === "stream" && !model.stream)
-    return Effect.fail(route.unsupported("media.stream", `${id} does not stream; use Transcription.generate`))
  if (request.diarize === true && !model.diarize)
    return Effect.fail(route.unsupported("media.diarize", `${id} does not diarize; use gpt-4o-transcribe-diarize`))
  if (request.prompt !== undefined && model.diarize)
@@ -173,7 +188,7 @@ const fromRequest = Effect.fn("OpenAITranscription.fromRequest")(function* (requ
      timestamp_granularities: responseFormat === "verbose_json" ? [request.timestamps] : undefined,
      // Diarizing audio longer than 30 seconds requires a chunking strategy.
      chunking_strategy: model.diarize ? "auto" : undefined,
-      stream: request.mode === "stream" ? true : undefined,
+      stream: streamsEvents(request) ? true : undefined,
    },
    {
      overlay: mergeJsonRecords(request.providerOptions, request.http?.body),
@@ -196,7 +211,15 @@ const segment = (value: Schema.Schema.Type<typeof Segment>): TranscriptionSegmen
 })

 const onEvent = Effect.fn("OpenAITranscription.onEvent")(function* (state: State, frame: string) {
+  if (!EVENT_TYPES.has((yield* decodeEventType(frame)).type)) return [state, []] as const
  const event = yield* decodeEvent(frame)
+  if (event.type === "error")
+    return yield* new AIError({
+      reason: classifyProviderFailure({
+        message: `${route.name} stream failed: ${event.message ?? event.error?.message ?? "unknown error"}`,
+        rawBody: frame,
+      }),
+    })
  if (event.type === "transcript.text.done") return [{ ...state, transcript: event }, []] as const
  if (event.type === "transcript.text.delta")
    return [state, event.delta.length === 0 ? [] : [TranscriptionTextDeltaEvent.make({ delta: event.delta })]] as const
@@ -246,7 +269,7 @@ export const protocol = MediaProtocol.stream<Request, TranscriptionEvent, Frame,
  unsupported: ["speakers"],
  body: { from: fromRequest },
  frames: (bytes, context) =>
-    context.request.mode === "stream"
+    streamsEvents(context.request)
      ? Framing.sse.frame(bytes)
      : Framing.document.frame(bytes).pipe(Stream.map((document) => ({ document }))),
  initial: () => ({ segments: [] }),
@@ -132,8 +132,7 @@ const decodeResult = Effect.fn("ReplicateImages.decodeResult")(function* (
      status,
      `${route.name} prediction ${context.token.id} ${prediction.status}${typeof prediction.error === "string" ? `: ${prediction.error}` : ""}`,
    )
-  if (status !== "completed")
-    return yield* output.invalid(`${route.name} prediction ${context.token.id} has not finished`)
+  if (status !== "completed") return yield* output.pending(context.token.id)
  if (prediction.data_removed === true)
    return yield* output.ended("expired", `${route.name} removed the output of prediction ${context.token.id}`)
  if (!isOutput(prediction.output))
@@ -137,12 +137,16 @@ const decodeResult = Effect.fn("RunwayVideo.decodeResult")(function* (
    const message = `${route.name} task failed${code === undefined ? "" : ` (${code})`}${task.failure ? `: ${task.failure}` : ""}`
    // Runway failure codes are dotted paths; every moderation outcome carries a SAFETY segment.
    if (code !== undefined && /(^|\.)SAFETY(\.|$)/.test(code)) return yield* output.contentPolicy(message)
-    return yield* output.ended("failed", message)
+    // ASSET.INVALID rejects the caller's input media; Runway documents it as not retryable.
+    return yield* output.ended(
+      "failed",
+      message,
+      code !== undefined && /^ASSET\.INVALID(\.|$)/.test(code) ? "InvalidRequest" : "ProviderInternal",
+    )
  }
  if (status === "cancelled")
    return yield* output.ended("cancelled", `${route.name} task ${context.token.taskID} was cancelled`)
-  if (status !== "completed")
-    return yield* output.invalid(`${route.name} task ${context.token.taskID} has not finished`)
+  if (status !== "completed") return yield* output.pending(context.token.taskID)
  const urls = task.output ?? []
  if (urls.length === 0) return yield* output.invalid(`${route.name} task succeeded without any output`)
  return new VideoResponse({
@@ -171,7 +175,7 @@ export const protocol = MediaProtocol.queued<Request, VideoResponse, Token>(rout
  start: { body: { from: fromRequest }, decode: decodeStart },
  status: { path: taskPath, decode: decodeStatus },
  result: { path: taskPath, decode: decodeResult },
-  cancel: { method: "DELETE", path: taskPath },
+  cancel: { method: "DELETE", path: taskPath, activeOnly: true },
 })

 const startPath = (request: Request) => {
@@ -175,7 +175,7 @@ const decodeUpscaleResult = Effect.fn("StabilityImages.decodeUpscaleResult")(fun
 ) {
  if (response.status === 202) {
    const output = yield* upscaleRoute.text(response)
-    return yield* output.invalid(`${upscaleRoute.name} upscale ${context.token.id} has not finished`)
+    return yield* output.pending(context.token.id)
  }
  return yield* decodeUpscaleImage(response)
 })
@@ -1,4 +1,5 @@
 // Shared counter and TTL mapping for provider cache-marker lowering.
+import type { CacheHint } from "../../schema/index.js"

 export interface Breakpoints {
  remaining: number
@@ -11,3 +12,12 @@ export const newBreakpoints = (cap: number): Breakpoints => ({ remaining: cap, d
 // requests omit the wire TTL and use the provider default.
 export const ttlBucket = (ttlSeconds: number | undefined): "1h" | undefined =>
  ttlSeconds !== undefined && ttlSeconds >= 3600 ? "1h" : undefined
+
+export const cacheControl = () => {
+  const breakpoints = newBreakpoints(4)
+  return (cache: CacheHint | undefined) => {
+    if (cache === undefined || breakpoints.remaining === 0) return undefined
+    breakpoints.remaining -= 1
+    return { type: "ephemeral" as const, ttl: ttlBucket(cache.ttlSeconds) }
+  }
+}
@@ -0,0 +1,19 @@
+import type { LLMRequest } from "../../schema/index.js"
+
+export const THINKING_BINDING_BETA = "thinking-binding-controls-2026-08-01"
+
+// Accept gateway namespaces and Vertex suffixes without treating a snapshot date as a minor version.
+export const claudeVersion = (id: string) => {
+  const match = /(?:^|[./])claude-(?<family>[a-z]+)-(?<major>\d+)(?:[.-](?<minor>\d{1,2}))?(?:$|[-:@])/.exec(
+    id.toLowerCase(),
+  )?.groups
+  if (!match) return undefined
+  return { family: match.family, major: Number(match.major), minor: Number(match.minor ?? 0) }
+}
+
+export const supportsThinkingBlockBinding = (model: LLMRequest["model"]) => {
+  const override = model.compatibility?.supportsThinkingBlockBinding
+  if (override !== undefined) return override
+  const version = claudeVersion(model.id)
+  return version !== undefined && (version.major > 5 || (version.major === 5 && version.minor >= 1))
+}
@@ -72,6 +72,44 @@ export const blocked = (name: string, chunk: Chunk, frame: string) => {
  })
 }

+const CONTENT_FILTER_REASONS = new Set([
+  "IMAGE_SAFETY",
+  "RECITATION",
+  "SAFETY",
+  "BLOCKLIST",
+  "PROHIBITED_CONTENT",
+  "SPII",
+  "MODEL_ARMOR",
+  "IMAGE_PROHIBITED_CONTENT",
+  "IMAGE_RECITATION",
+  "LANGUAGE",
+])
+
+/** Finish reasons for which Gemini stops output on safety or policy grounds. */
+export const contentFiltered = (finishReason: string | undefined) =>
+  finishReason !== undefined && CONTENT_FILTER_REASONS.has(finishReason)
+
+/** Callers check that the response produced no output: a policy stop after output is a partial result instead. */
+export const withheld = (name: string, chunk: Chunk, frame: string) => {
+  const finishReason = chunk.candidates?.[0]?.finishReason
+  if (!contentFiltered(finishReason)) return undefined
+  return new AIError({
+    reason: new ContentPolicyError({ message: `${name} withheld its output (${finishReason})`, body: frame }),
+  })
+}
+
+/** Any finish reason other than `STOP` means the output may be cut short, so it is surfaced rather than dropped. */
+export const notices = (name: string, state: Metadata): ReadonlyArray<Media.Notice> | undefined =>
+  state.finishReason === undefined || state.finishReason === "STOP"
+    ? undefined
+    : [
+        {
+          type: contentFiltered(state.finishReason) ? "filtered" : "other",
+          message: `${name} finished with ${state.finishReason}`,
+          providerMetadata: { google: { finishReason: state.finishReason } },
+        },
+      ]
+
 export const usage = (usage: UsageMetadata | undefined): MediaUsage | undefined =>
  usage === undefined
    ? undefined
@@ -71,8 +71,9 @@ export const imageOutput = (
 }

 /**
- * Append multipart text fields: strings as-is, other values as JSON, or arrays as repeated parts named `key[]` or
- * `key` with `repeatArrays`. `overlay` keys in `reserved` are dropped so `http.body` cannot replace route-owned fields.
+ * Append multipart text fields: strings as-is, other values as JSON, or scalar arrays as one part per item with
+ * `repeatArrays`, named `key[]` or `key`. `overlay` keys in `reserved` are dropped so `http.body` cannot replace
+ * route-owned fields.
 */
 export const appendFields = (
  form: FormData,
@@ -85,7 +86,7 @@ export const appendFields = (
 ) => {
  const overlay = Object.entries(options.overlay ?? {}).filter(([key]) => !options.reserved.has(key))
  Object.entries(mergeJsonRecords(fields, Object.fromEntries(overlay)) ?? {}).forEach(([key, value]) => {
-    if (Array.isArray(value) && options.repeatArrays !== undefined)
+    if (Array.isArray(value) && value.every(isScalar) && options.repeatArrays !== undefined)
      return value.forEach((item) => form.append(options.repeatArrays === "key[]" ? `${key}[]` : key, String(item)))
    form.append(key, typeof value === "string" ? value : encodeJson(value))
  })
@@ -0,0 +1,10 @@
+/** Split an ordered token list into runs of consecutive tokens with the same speaker. */
+export const group = <Item>(items: ReadonlyArray<Item>, speaker: (item: Item) => unknown) =>
+  items.reduce<Array<Array<Item>>>((turns, item) => {
+    const last = turns.at(-1)
+    if (last === undefined || speaker(last[0]) !== speaker(item)) return [...turns, [item]]
+    last.push(item)
+    return turns
+  }, [])
+
+export * as SpeakerTurns from "./speaker-turns.js"
@@ -85,6 +85,7 @@ export const finish = (
    readonly mediaType: string | undefined
    readonly info?: Media.Info
    readonly usage?: MediaUsage
+    readonly notices?: ReadonlyArray<Media.Notice>
    readonly providerMetadata?: ProviderMetadata
    readonly detail?: string
  },
@@ -97,6 +98,7 @@ export const finish = (
    SpeechFinishEvent.make({
      audio: Media.bytes(concatBytes(state.chunks), output.mediaType, { info: output.info }),
      usage: output.usage,
+      notices: output.notices,
      providerMetadata: output.providerMetadata,
    }),
  ])
@@ -147,7 +147,12 @@ export const appendOrStart = <K extends StreamKey>(
  route: string,
  tools: State<K>,
  key: K,
-  delta: { readonly id?: string; readonly name?: string; readonly text: string },
+  delta: {
+    readonly id?: string
+    readonly name?: string
+    readonly text: string
+    readonly providerMetadata?: ProviderMetadata
+  },
  missingToolMessage: string,
 ): AppendOutcome<K> | AIError => {
  const current = tools[key]
@@ -161,7 +166,7 @@ export const appendOrStart = <K extends StreamKey>(
    namespace: current?.namespace,
    input: `${current?.input ?? ""}${delta.text}`,
    providerExecuted: current?.providerExecuted,
-    providerMetadata: current?.providerMetadata,
+    providerMetadata: current?.providerMetadata ?? delta.providerMetadata,
  }
  if (current && delta.text.length === 0 && current.id === id && current.name === name)
    return { tools, tool: current, events: [] }
@@ -101,7 +101,8 @@ const decodeResponse = Effect.fn("XAIImages.decodeResponse")(function* (
  )
  if (images.length === 0) return yield* output.invalid(`${route.name} returned no images`)
  const usage = ProviderShared.isRecord(decoded.usage) ? decoded.usage : undefined
-  // xAI reports image counts rather than tokens, seconds, or credits; the raw record stays in provider metadata.
+  // xAI reports a USD cost (`cost_in_usd_ticks`) rather than tokens, seconds, or credits; the raw record stays in
+  // provider metadata.
  return new ImageResponse({
    images,
    providerMetadata: usage === undefined ? undefined : { xai: { usage } },
@@ -1,112 +0,0 @@
-import { Effect } from "effect"
-import { MediaProtocol } from "../route/media-protocol.js"
-import { MediaRoute } from "../route/media.js"
-import { mergeJsonRecords } from "../schema/index.js"
-import { SpeechModel, type SpeechEvent, type SpeechRequestFor } from "../speech.js"
-import { SpeechStream } from "./utils/speech-stream.js"
-
-const route = MediaProtocol.identity({ id: "xai-speech", name: "xAI Speech", provider: "xai" })
-export const DEFAULT_BASE_URL = "https://api.x.ai/v1"
-export const PATH = "/tts"
-const DEFAULT_SAMPLE_RATE = 24000
-
-// ---------------------------------------------------------------------------
-// 1. Public model input
-// ---------------------------------------------------------------------------
-
-/** `voice`, `format`, `speed`, and `language` are common request fields; other native body fields pass through. */
-export type XAISpeechOptions = {
-  readonly sampleRate?: 8000 | 16000 | 22050 | 24000 | 44100 | 48000
-  /** MP3 only. */
-  readonly bitRate?: 32000 | 64000 | 96000 | 128000 | 192000
-  readonly optimize_streaming_latency?: number
-  readonly text_normalization?: boolean
-  readonly replace?: Readonly<Record<string, string>>
-} & Record<string, unknown>
-
-export type Request = SpeechRequestFor<XAISpeechOptions>
-
-// ---------------------------------------------------------------------------
-// 4. Parser state
-// ---------------------------------------------------------------------------
-
-type State = SpeechStream.Audio
-
-// ---------------------------------------------------------------------------
-// 5. Request body construction
-// ---------------------------------------------------------------------------
-
-/** `output_format.codec` values; headerless codecs map to the PCM encoding of their samples. */
-const CODECS = new Map<string, SpeechStream.PcmEncoding | undefined>([
-  ["mp3", undefined],
-  ["wav", undefined],
-  ["pcm", "pcm_s16le"],
-  ["mulaw", "pcm_mulaw"],
-  ["alaw", "pcm_alaw"],
-])
-
-const outputFormat = Effect.fn("XAISpeech.outputFormat")(function* (request: Request) {
-  const codec = request.format ?? "mp3"
-  if (!CODECS.has(codec))
-    return yield* route.unsupported(
-      "media.format",
-      `${route.name} supports the mp3, wav, pcm, mulaw, and alaw formats, not "${codec}"`,
-    )
-  return { codec, sample_rate: request.providerOptions?.sampleRate, bit_rate: request.providerOptions?.bitRate }
-})
-
-// The TTS API has no model field, so the selected model id only names the model.
-const fromRequest = Effect.fn("XAISpeech.fromRequest")(function* (request: MediaProtocol.Addressed<Request>) {
-  const { sampleRate: _sampleRate, bitRate: _bitRate, ...native } = request.providerOptions ?? {}
-  return MediaProtocol.json(
-    mergeJsonRecords(
-      {
-        text: request.text,
-        voice_id: SpeechStream.voiceID(request.voice),
-        // `language` is required; `auto` detects it from the text.
-        language: request.language ?? "auto",
-        output_format: yield* outputFormat(request),
-        speed: request.speed,
-      },
-      native,
-      request.http?.body,
-    ) ?? {},
-  )
-})
-
-// ---------------------------------------------------------------------------
-// 6. Stream parsing
-// ---------------------------------------------------------------------------
-
-const finish = Effect.fn("XAISpeech.finish")(function* (state: State, context: MediaProtocol.ResponseContext<Request>) {
-  const format = yield* outputFormat(context.request)
-  const sampleRate = format.sample_rate ?? DEFAULT_SAMPLE_RATE
-  const encoding = CODECS.get(format.codec)
-  return yield* SpeechStream.finish(
-    route,
-    state,
-    encoding === undefined ? SpeechStream.container(format.codec, sampleRate) : SpeechStream.pcm(encoding, sampleRate),
-  )
-})
-
-// ---------------------------------------------------------------------------
-// 7. Protocol and route
-// ---------------------------------------------------------------------------
-
-/** The response body is the raw audio in both modes, so `stream` forwards body chunks as they arrive. */
-export const protocol = MediaProtocol.stream<Request, SpeechEvent, Uint8Array, State>(route, {
-  unsupported: ["instructions", "timestamps"],
-  body: { from: fromRequest },
-  frames: (bytes) => bytes,
-  initial: () => ({ chunks: [] }),
-  step: (state, frame) => Effect.succeed(SpeechStream.delta(state, frame)),
-  finish,
-})
-
-export const model = (input: MediaRoute.ModelInput) =>
-  SpeechModel.fromRoute<XAISpeechOptions, Uint8Array, State>({ protocol, baseURL: DEFAULT_BASE_URL, path: PATH }, input)
-
-export const XAISpeech = {
-  protocol,
-  model,
-} as const
@@ -1,179 +0,0 @@
-import { Effect, Schema } from "effect"
-import type { HttpClientResponse } from "effect/unstable/http"
-import { MediaProtocol } from "../route/media-protocol.js"
-import { MediaRoute } from "../route/media.js"
-import { mergeJsonRecords, type OpenString } from "../schema/index.js"
-import {
-  TranscriptionModel,
-  TranscriptionResponse,
-  type TranscriptionRequestFor,
-  type TranscriptionSegment,
-  type TranscriptionWord,
-} from "../transcription.js"
-import { mediaTypeExtension } from "../utils/media-type.js"
-import { ProviderShared } from "./shared.js"
-import { MediaInput } from "./utils/media-input.js"
-
-const route = MediaProtocol.identity({ id: "xai-transcription", name: "xAI Transcription", provider: "xai" })
-export const DEFAULT_BASE_URL = "https://api.x.ai/v1"
-export const PATH = "/stt"
-
-// ---------------------------------------------------------------------------
-// 1. Public model input
-// ---------------------------------------------------------------------------
-
-export type XAITranscriptionOptions = {
-  /** Inverse text normalization ("one hundred dollars" → "$100"); requires `language`. */
-  readonly format?: boolean
-  readonly keyterm?: ReadonlyArray<string>
-  readonly filler_words?: boolean
-  /** Headerless audio only; derived from `audio.info.encoding` and `audio.info.sampleRate` when omitted. */
-  readonly audio_format?: OpenString<"pcm" | "mulaw" | "alaw">
-  readonly sample_rate?: 8000 | 16000 | 22050 | 24000 | 44100 | 48000
-  readonly multichannel?: boolean
-  readonly channels?: number
-  readonly vad_threshold?: number
-} & Record<string, unknown>
-
-export type Request = TranscriptionRequestFor<XAITranscriptionOptions>
-
-// ---------------------------------------------------------------------------
-// 2. Response schema
-// ---------------------------------------------------------------------------
-
-const Word = Schema.Struct({
-  text: Schema.String,
-  start: Schema.Number,
-  end: Schema.Number,
-  confidence: Schema.optional(Schema.Number),
-  speaker: Schema.optional(Schema.Number),
-})
-
-const SttResponse = Schema.Struct({
-  text: Schema.String,
-  language: Schema.optional(Schema.String),
-  duration: Schema.optional(Schema.Number),
-  words: Schema.optional(Schema.Array(Word)),
-  channels: Schema.optional(
-    Schema.Array(
-      Schema.Struct({
-        index: Schema.Number,
-        text: Schema.String,
-        language: Schema.optional(Schema.String),
-        words: Schema.optional(Schema.Array(Word)),
-      }),
-    ),
-  ),
-})
-
-// ---------------------------------------------------------------------------
-// 5. Request body construction
-// ---------------------------------------------------------------------------
-
-/** xAI returns only words, so segments are diarized speaker turns, as with AssemblyAI utterances. */
-const wantsSegments = (request: Request) => request.diarize === true || request.timestamps === "segment"
-
-const RAW_AUDIO_FORMATS: Readonly<Record<string, string>> = {
-  pcm_s16le: "pcm",
-  pcm_mulaw: "mulaw",
-  pcm_alaw: "alaw",
-}
-
-const RESERVED_FORM_FIELDS = new Set(["file", "url", "model", "language", "diarize"])
-
-const fromRequest = Effect.fn("XAITranscription.fromRequest")(function* (request: Request) {
-  const audioFormat = RAW_AUDIO_FORMATS[request.audio.info?.encoding ?? ""]
-  const form = new FormData()
-  MediaInput.appendFields(
-    form,
-    {
-      model: request.model.id,
-      language: request.language,
-      diarize: wantsSegments(request) ? true : undefined,
-      audio_format: audioFormat,
-      sample_rate: audioFormat === undefined ? undefined : request.audio.info?.sampleRate,
-    },
-    {
-      overlay: mergeJsonRecords(request.providerOptions, request.http?.body),
-      reserved: RESERVED_FORM_FIELDS,
-      repeatArrays: "key",
-    },
-  )
-  // `file` must be the last field: options after it may be ignored for streamed uploads.
-  const url = ProviderShared.mediaUrl(request.audio)
-  if (url !== undefined) {
-    form.append("url", url)
-    return MediaProtocol.multipart(form)
-  }
-  const audio = yield* MediaInput.inlineBytes(route.id, request.audio)
-  const extension = mediaTypeExtension(request.audio.mediaType)
-  form.append(
-    "file",
-    MediaInput.blob(audio, request.audio.mediaType),
-    extension === undefined ? "audio" : `audio.${extension}`,
-  )
-  return MediaProtocol.multipart(form)
-})
-
-// ---------------------------------------------------------------------------
-// 6. Response decoding
-// ---------------------------------------------------------------------------
-
-const decodeStt = route.decodeJson(SttResponse)
-
-const word = (value: typeof Word.Type): TranscriptionWord => ({
-  text: value.text,
-  startSeconds: value.start,
-  endSeconds: value.end,
-  speaker: value.speaker === undefined ? undefined : String(value.speaker),
-  confidence: value.confidence,
-})
-
-const speakerSegments = (words: ReadonlyArray<TranscriptionWord>) =>
-  words.reduce<Array<TranscriptionSegment>>((turns, next) => {
-    const last = turns.at(-1)
-    if (last === undefined || last.speaker !== next.speaker)
-      return [
-        ...turns,
-        { text: next.text, startSeconds: next.startSeconds, endSeconds: next.endSeconds, speaker: next.speaker },
-      ]
-    turns[turns.length - 1] = { ...last, text: `${last.text} ${next.text}`, endSeconds: next.endSeconds }
-    return turns
-  }, [])
-
-const decodeResponse = Effect.fn("XAITranscription.decodeResponse")(function* (
-  response: HttpClientResponse.HttpClientResponse,
-  context: MediaProtocol.DecodeContext<Request>,
-) {
-  const output = yield* decodeStt(response)
-  const transcript = output.value
-  const words = transcript.words?.map(word)
-  const duration = transcript.duration
-  return new TranscriptionResponse({
-    text: transcript.text,
-    segments: words === undefined || !wantsSegments(context.request) ? undefined : speakerSegments(words),
-    words,
-    language: transcript.language?.toLowerCase(),
-    durationSeconds: duration,
-    usage: duration === undefined ? undefined : { type: "seconds", seconds: duration },
-    providerMetadata: transcript.channels === undefined ? undefined : { xai: { channels: transcript.channels } },
-  })
-})
-
-// ---------------------------------------------------------------------------
-// 7. Protocol and route
-// ---------------------------------------------------------------------------
-
-export const protocol = MediaProtocol.inline<Request, TranscriptionResponse>(route, {
-  unsupported: ["prompt", "speakers"],
-  body: { from: fromRequest },
-  response: { decode: decodeResponse },
-})
-
-export const model = (input: MediaRoute.ModelInput) =>
-  TranscriptionModel.fromRoute<XAITranscriptionOptions>({ protocol, baseURL: DEFAULT_BASE_URL, path: PATH }, input)
-
-export const XAITranscription = {
-  protocol,
-  model,
-} as const
@@ -65,6 +65,13 @@ const STATUS = {
  expired: "expired",
 } as const satisfies Record<string, Status>

+// Documented video error codes; `service_unavailable`, `internal_error`, and unknown codes are provider-side.
+const FAILURE = {
+  invalid_argument: "InvalidRequest",
+  failed_precondition: "InvalidRequest",
+  permission_denied: "Authentication",
+} as const satisfies Record<string, MediaProtocol.Failure>
+
 // ---------------------------------------------------------------------------
 // 5. Request body construction
 // ---------------------------------------------------------------------------
@@ -136,14 +143,14 @@ const decodeResult = Effect.fn("XAIVideo.decodeResult")(function* (
  const output = yield* decodeVideoStatus(response)
  const decoded = output.value
  const status = yield* MediaProtocol.status(STATUS, decoded.status, output)
-  if (status === "running")
-    return yield* output.invalid(`${route.name} request ${context.token.requestID} has not finished`)
+  if (status === "running") return yield* output.pending(context.token.requestID)
  if (status === "failed") {
    const code = decoded.error?.code ?? undefined
    const message = decoded.error?.message ?? undefined
    return yield* output.ended(
      "failed",
      `${route.name} generation failed${code === undefined ? "" : ` (${code})`}${message === undefined ? "" : `: ${message}`}`,
+      MediaProtocol.failure(FAILURE, code),
    )
  }
  if (status !== "completed")
@@ -41,16 +41,20 @@ const Body = Schema.Struct({
  user_id: Options.fields.userID,
 })

+// Tool streaming was introduced in GLM-4.6; later versions inherit support.
+const supportsToolStreaming = (modelID: string) => {
+  const match = /(?:^|\/)glm-(\d+)(?:\.(\d+))?(?:-|$)/i.exec(modelID)
+  return match !== null && (Number(match[1]) > 4 || (Number(match[1]) === 4 && Number(match[2] ?? 0) >= 6))
+}
+
 const fromRequest = Effect.fn("ZAIChat.fromRequest")(function* (request: LLMRequest) {
  const options = yield* ProviderShared.validateWith(Schema.decodeUnknownEffect(Options))(request.providerOptions ?? {})
  const body = yield* OpenAIChat.protocol.body.from(request)
  return {
    ...body,
    thinking: options.thinking,
-    // Tool streaming was introduced in GLM-4.6; older models must not receive the opt-in.
    tool_stream:
-      options.toolStream ??
-      (body.tools?.length && /^glm-(?:4\.[67]|5(?:[.-]|$))/i.test(request.model.id) ? true : undefined),
+      options.toolStream ?? (body.tools?.length && supportsToolStreaming(request.model.id) ? true : undefined),
    do_sample: options.doSample,
    response_format: options.responseFormat,
    request_id: options.requestID,
@@ -1,7 +1,6 @@
-import { Effect, Schema } from "effect"
+import { Duration, Effect, Schema } from "effect"
 import type { HttpClientResponse } from "effect/unstable/http"
 import { ImageModel, ImageResponse, type ImageRequestFor } from "../image.js"
-import { Media } from "../media.js"
 import { MediaProtocol } from "../route/media-protocol.js"
 import { MediaRoute } from "../route/media.js"
 import { mergeJsonRecords, type OpenString } from "../schema/index.js"
@@ -9,6 +8,7 @@ import { mergeJsonRecords, type OpenString } from "../schema/index.js"
 const route = MediaProtocol.identity({ id: "zai-images", name: "Z.ai Images", provider: "zai" })
 export const DEFAULT_BASE_URL = "https://api.z.ai/api/paas/v4"
 export const PATH = "/images/generations"
+const OUTPUT_RETENTION = Duration.days(30)

 // ---------------------------------------------------------------------------
 // 1. Public model input
@@ -76,7 +76,7 @@ const decodeResponse = Effect.fn("ZAIImages.decodeResponse")(function* (
  const filters = decoded.content_filter ?? []
  return new ImageResponse({
    // Z.ai returns only URLs and no content type; the media type resolves when the asset is materialized.
-    images: decoded.data.map((item) => Media.url(item.url)),
+    images: yield* Effect.forEach(decoded.data, (item) => MediaProtocol.expiringUrl(item.url, OUTPUT_RETENTION)),
    // Z.ai reports applied content filters alongside a successful result; surface them instead of dropping them.
    notices:
      filters.length === 0
@@ -1,4 +1,4 @@
-import { Option, Schema } from "effect"
+import { Option, Schema, SchemaGetter } from "effect"
 import {
  AuthenticationError,
  ContentPolicyError,
@@ -16,24 +16,34 @@ import {
 const patterns = [
  /prompt is too long/i,
  /input is too long for requested model/i,
+  // Cloudflare Workers AI reports this as HTTP 413.
+  /exceeded this model context window limit/i,
  /exceeds the context window/i,
  /exceeds (?:the )?(?:model'?s )?maximum context length(?: of [\d,]+ tokens?|\s*\([\d,]+\))/i,
  /input token count.*exceeds the maximum/i,
+  // Amazon Nova on Bedrock reports this as a mid-stream validationException.
+  /number of input tokens exceeds maximum length/i,
  /tokens in request more than max tokens allowed/i,
  /maximum prompt length is \d+/i,
  /reduce the length of the messages/i,
+  // DeepInfra
+  /requested input length \d+ exceeds maximum input length/i,
  /maximum context length is \d+ tokens/i,
  /exceeds (?:the )?maximum allowed input length of [\d,]+ tokens?/i,
-  /input \(\d+ tokens\) is longer than the model'?s context length \(\d+ tokens\)/i,
+  // Novita omits the token counts.
+  /input(?: \(\d+ tokens\))? is longer than the model'?s context length/i,
  /exceeds the limit of \d+/i,
  /exceeds the available context size/i,
  /greater than the context length/i,
+  // Hugging Face Text Generation Inference, e.g. Together
+  /`inputs` tokens \+ `max_new_tokens` must be <= \d+/i,
  /context window exceeds limit/i,
  /exceeded model token limit/i,
  /context[_ ]length[_ ]exceeded/i,
  /context length is only \d+ tokens/i,
  /input length.*exceeds.*context length/i,
-  /prompt too long; exceeded (?:max )?context length/i,
+  // Z.ai code 1261 arrives as `Prompt too long` or `Prompt 超长`.
+  /prompt (?:too long|超长)/i,
  /too large for model with \d+ maximum context length/i,
  /prompt has [\d,]+ tokens?, but the configured context size is [\d,]+ tokens?/i,
  /model_context_window_exceeded/i,
@@ -45,7 +55,13 @@ const patterns = [

 const payloadPatterns = [/request entity too large/i, /payload too large/i, /request too large/i]

-const exclusions = [/^(throttling error|service unavailable):/i, /rate limit/i, /too many requests/i]
+const exclusions = [
+  /^(throttling error|service unavailable):/i,
+  /rate limit/i,
+  /too many requests/i,
+  // Cohere reports an output limit above the model maximum as "too many tokens"; compaction cannot fix it.
+  /max[_ ]tokens must be less than/i,
+]

 export const isContextOverflow = (message: string) =>
  !exclusions.some((pattern) => pattern.test(message)) &&
@@ -58,6 +74,47 @@ export const isContextOverflowFailure = (failure: unknown) =>
    ? failure.reason._tag === "InvalidRequest" && failure.reason.classification === "context-overflow"
    : Schema.is(ProviderErrorEvent)(failure) && failure.classification === "context-overflow"

+/**
+ * Whether a failed call may succeed when sent again: rate limits, provider-side failures, transport failures that did
+ * not deliver an accepted write, and unrecognized failures. Callers decide which calls are safe to repeat.
+ */
+export const isRetryable = (error: AIError) => {
+  const override = error.reason.http?.headers["x-should-retry"]
+  if (override === "true") return true
+  if (override === "false") return false
+  switch (error.reason._tag) {
+    case "RateLimit":
+    case "ProviderInternal":
+      return true
+    // A WebSocket acknowledgment marks delivery accepted before model output may exist.
+    // Read failures can still recover; the caller chooses retry versus continuation from durable output.
+    case "Transport":
+      return (
+        error.reason.delivery !== "rejected" &&
+        (error.reason.delivery !== "accepted" || error.reason.operation === "read")
+      )
+    case "InvalidProviderOutput":
+      return error.reason.classification === "incomplete-stream"
+    // Unrecognized failures retry: classification records affirmative
+    // deterministic evidence, and transient failures are exactly the ones
+    // that arrive in shapes no classifier anticipates.
+    case "UnknownProvider":
+      return true
+    case "Authentication":
+    case "QuotaExceeded":
+    case "ContentPolicy":
+    case "InvalidRequest":
+    case "UnsupportedOperation":
+    case "NoRoute":
+    case "Timeout":
+      return false
+    default: {
+      const exhaustive: never = error.reason
+      return exhaustive
+    }
+  }
+}
+
 const decodeJson = Schema.decodeUnknownOption(Schema.fromJsonString(Schema.Unknown))
 // OpenCode Zen reports account caps as typed 429/402 errors that are not throttles.
 const QUOTA_CODES = new Set([
@@ -68,7 +125,9 @@ const QUOTA_CODES = new Set([
  "freeusagelimiterror",
  "creditlimitexceeded",
 ])
-const AUTH_CODES = new Set(["authentication_error", "permission_error"])
+// Google reports an invalid API key as HTTP 400 INVALID_ARGUMENT with this `details[].reason`.
+// Z.ai's Responses API reports account and plan rejections mid-stream as `permission_denied`.
+const AUTH_CODES = new Set(["authentication_error", "permission_error", "permission_denied", "api_key_invalid"])
 const SERVER_CODES = new Set([
  "api_error",
  "internal_error",
@@ -82,6 +141,7 @@ const SERVER_CODES = new Set([
 ])
 // `invalid_request` is the Vercel AI Gateway's code for an upstream request rejection.
 const INVALID_REQUEST_CODES = new Set([
+  "model_not_found",
  "invalid_prompt",
  "invalid_request",
  "invalid_request_error",
@@ -101,18 +161,57 @@ const CONTENT_POLICY_CODES = new Set([
 // OpenCode Zen replaces upstream codes outside its allow-list but keeps the original
 // as a `[code]` label at the start of the rewritten message.
 const GATEWAY_CODE_LABEL = /^[^:\n]+: \[([A-Za-z0-9_.-]+)\]/
+// xAI reports an invalid API key as HTTP 400 with the generic `invalid-argument` code.
+const AUTH_TEXT = /incorrect api key provided/i
 const RATE_LIMIT_TEXT = /rate increased too quickly|rate[-_\s]?limit|too[_\s]?many[_\s]?requests/i
 // Only consulted on 429, where throttles and account caps share a status.
-const QUOTA_TEXT = /insufficient[-_\s]?quota|quota[-_\s]?exceeded|budget exceeded|usage limit/i
+// Z.ai reports balance, plan expiry, plan limits, and plan model access on 429.
+const QUOTA_TEXT =
+  /insufficient[-_\s]?(?:quota|balance)|quota[-_\s]?exceeded|budget exceeded|usage limit|limit exhausted|package has expired|plan does not yet include/i
 // Policy rejections without a dedicated code, matched against the provider's own
 // explanation only. OpenAI reuses `invalid_prompt` for usage-policy rejections while
 // Bedrock Mantle reuses it for schema validation; Anthropic reports blocked output
 // under `invalid_request_error`.
 const CONTENT_POLICY_TEXT =
-  /violating our usage policy|blocked by content filtering policy|content[-_\s]?policy|rejected as a result of our safety system/i
+  /violating our usage policy|blocked by content filtering policy|content[-_\s]?policy|rejected as a result of our safety system|detected potentially unsafe or sensitive content/i
 const SERVER_ERROR_TEXT =
  /\b(?:try again|(?:please |you can )?retry (?:the |this |your )?request|try (?:the |this |your )?request again|(?:currently |temporarily )?at capacity|overloaded|temporarily unavailable|service[-_\s]?unavailable|(?:server|internal)[-_\s]?error|server (?:is )?busy|provider returned (?:an )?error|resource[-_\s]?exhausted|upstream (?:connect|connection|request)|request buffer limit while retrying upstream)\b/i

+const Message = Schema.String.check(Schema.isPattern(/\S/))
+
+const messageAt = <Fields extends Schema.Struct.Fields>(
+  fields: Fields,
+  message: (body: Schema.Struct<Fields>["Type"]) => string,
+) =>
+  Schema.Struct(fields).pipe(
+    Schema.decodeTo(Schema.String, {
+      decode: SchemaGetter.transform(message),
+      encode: SchemaGetter.forbidden(() => "Provider error messages are decode-only"),
+    }),
+  )
+
+// Common error body layouts that carry a human-readable message, in priority order.
+// Provider-specific layouts belong in their protocol.
+const decodeMessage = Schema.decodeUnknownOption(
+  Schema.fromJsonString(
+    Schema.Union([
+      messageAt({ error: Schema.Struct({ message: Message }) }, (body) => body.error.message),
+      messageAt({ error: Message }, (body) => body.error),
+      messageAt({ message: Message }, (body) => body.message),
+      // AWS services
+      messageAt({ Message: Message }, (body) => body.Message),
+      // RFC 9457 problem details
+      messageAt({ detail: Message }, (body) => body.detail),
+      messageAt(
+        { errors: Schema.NonEmptyArray(Schema.Struct({ message: Message })) },
+        (body) => body.errors[0].message,
+      ),
+    ]),
+  ),
+)
+
+export const providerErrorMessage = (body: string) => Option.getOrUndefined(decodeMessage(body))
+
 export interface ProviderFailure {
  readonly message: string
  readonly status?: number | undefined
@@ -165,7 +264,12 @@ export function classifyProviderFailure(input: ProviderFailure): AIError["reason
    (input.status === 429 && QUOTA_TEXT.test(text))
  )
    return new QuotaExceededError(details)
-  if (input.status === 401 || input.status === 403 || codes.some((code) => AUTH_CODES.has(code)))
+  if (
+    input.status === 401 ||
+    input.status === 403 ||
+    codes.some((code) => AUTH_CODES.has(code)) ||
+    (input.status === 400 && AUTH_TEXT.test(text))
+  )
    return new AuthenticationError(details)
  if (
    input.status === 429 ||
@@ -218,6 +322,10 @@ function providerCodes(value: unknown) {
    error?.type,
    error?.status,
    error?.error_type,
+    // Google `google.rpc.ErrorInfo` details carry the specific reason.
+    ...(Array.isArray(error?.details)
+      ? error.details.map((detail) => (isRecord(detail) ? detail.reason : undefined))
+      : []),
    inner?.code,
    metadata?.error_type,
    responseError?.code,
@@ -28,7 +28,8 @@ export type Settings = ProviderPackage.Settings &
    readonly provider?: string
  }

-export const routes = [AnthropicMessages.route]
+const compatibleRoute = AnthropicMessages.route.with({ id: "anthropic-compatible-messages", provider: id })
+export const routes = [compatibleRoute]

 const auth = (input: ProviderAuthOption<"optional">) => {
  if ("auth" in input && input.auth) return input.auth
@@ -43,7 +44,7 @@ export const configure = (input: Config) => {
      message: "Anthropic-compatible providers require a baseURL",
    })
  const { provider: _, baseURL, apiKey: _apiKey, auth: _auth, ...rest } = input
-  const route = AnthropicMessages.route.with({
+  const route = (provider === "anthropic" ? AnthropicMessages.route : compatibleRoute).with({
    ...rest,
    provider,
    endpoint: { baseURL },
@@ -42,14 +42,24 @@ export type Settings = ProviderPackage.Settings &

 export const baseURL = (input: GatewayURL) => {
  if (input.baseURL) return input.baseURL
-  if (!input.accountId)
-    throw new ProviderConfigurationError({
-      provider: id,
-      message: "CloudflareAIGateway.configure requires accountId unless baseURL is supplied",
-    })
-  return `https://api.cloudflare.com/client/v4/accounts/${encodeURIComponent(input.accountId)}/ai/v1`
+  return `https://api.cloudflare.com/client/v4/accounts/${encodeURIComponent(requireAccountId(input))}/ai/v1`
 }

+// Cloudflare's REST API rejects parts of the Anthropic Messages and OpenAI Responses request shapes, such as
+// system prompt blocks and `minimal` reasoning effort, so these models use the provider-native gateway endpoints.
+const passthroughURL = (input: GatewayURL & GatewayOptions, provider: "anthropic/v1" | "openai") =>
+  `https://gateway.ai.cloudflare.com/v1/${encodeURIComponent(requireAccountId(input))}/${encodeURIComponent(gatewayId(input))}/${provider}`
+
+const requireAccountId = (input: GatewayURL) => {
+  if (input.accountId) return input.accountId
+  throw new ProviderConfigurationError({
+    provider: id,
+    message: "CloudflareAIGateway.configure requires accountId unless baseURL is supplied",
+  })
+}
+
+const gatewayId = (input: GatewayOptions) => input.gatewayId?.trim() || "default"
+
 export const responsesRoute = OpenAIResponses.route.with({
  id: "cloudflare-ai-gateway-responses",
  provider: id,
@@ -70,16 +80,12 @@ export const route = OpenAIChat.route.with({

 export const routes = [responsesRoute, messagesRoute, route]

-const auth = (input: LanguageModelOptions) => {
-  if ("auth" in input && input.auth) return input.auth
-  return Auth.optional(input.gatewayApiKey ?? ("apiKey" in input ? input.apiKey : undefined), "apiKey")
+const credential = (input: LanguageModelOptions) =>
+  Auth.optional(input.gatewayApiKey ?? ("apiKey" in input ? input.apiKey : undefined), "apiKey")
    .orElse(Auth.config(authEnvVars[0]))
    .orElse(Auth.config(authEnvVars[1]))
-    .bearer()
-}

 const headers = (input: LanguageModelOptions) => ({
-  ...(input.gatewayId === undefined ? {} : { "cf-aig-gateway-id": input.gatewayId.trim() || "default" }),
  ...(input.metadata === undefined
    ? {}
    : { "cf-aig-metadata": Schema.encodeSync(Schema.fromJsonString(Schema.Unknown))(input.metadata) }),
@@ -90,31 +96,42 @@ const headers = (input: LanguageModelOptions) => ({
  ...input.headers,
 })

-const modelID = (input: string | ModelID) => {
-  const value = String(input)
-  if (value.startsWith("workers-ai/")) return value.slice("workers-ai/".length)
-  if (value.startsWith("anthropic/")) return `anthropic/${value.slice("anthropic/".length).replaceAll(".", "-")}`
-  return value
-}
-
 export const configure = (input: LanguageModelOptions) => {
-  const defaults = {
+  const custom = "auth" in input && input.auth ? input.auth : undefined
+  const rest = {
    endpoint: { baseURL: baseURL(input) },
-    auth: auth(input),
-    headers: headers(input),
+    auth: custom ?? credential(input).bearer(),
+    headers: { "cf-aig-gateway-id": gatewayId(input), ...headers(input) },
    http: input.http,
    providerOptions: input.providerOptions,
  }
-  const responses = responsesRoute.with(defaults)
-  const messages = messagesRoute.with(defaults)
-  const chat = route.with(defaults)
+  // A custom baseURL stands in for the REST API, so every model keeps using it with gateway model IDs.
+  const native = input.baseURL === undefined
+  const passthrough = (provider: "anthropic/v1" | "openai") =>
+    native
+      ? {
+          ...rest,
+          endpoint: { baseURL: passthroughURL(input, provider) },
+          auth: custom ?? Auth.bearerHeader("cf-aig-authorization", credential(input)),
+          headers: headers(input),
+        }
+      : rest
+  const responses = responsesRoute.with(passthrough("openai"))
+  const messages = messagesRoute.with(passthrough("anthropic/v1"))
+  const chat = route.with(rest)
  return {
    id,
    model: (input: string | ModelID) => {
-      const wire = modelID(input)
-      if (String(input).startsWith("openai/")) return responses.model<OpenAIProviderOptionsInput>({ id: wire })
-      if (String(input).startsWith("anthropic/")) return messages.model<OpenAIProviderOptionsInput>({ id: wire })
-      return chat.model<OpenAIProviderOptionsInput>({ id: wire })
+      const value = String(input)
+      if (value.startsWith("openai/"))
+        return responses.model<OpenAIProviderOptionsInput>({ id: native ? value.slice("openai/".length) : value })
+      if (value.startsWith("anthropic/"))
+        return messages.model<OpenAIProviderOptionsInput>({
+          id: native ? value.slice("anthropic/".length).replaceAll(".", "-") : value,
+        })
+      if (value.startsWith("workers-ai/"))
+        return chat.model<OpenAIProviderOptionsInput>({ id: value.slice("workers-ai/".length) })
+      return chat.model<OpenAIProviderOptionsInput>({ id: value })
    },
    configure,
  }
@@ -0,0 +1,66 @@
+import { CohereChat } from "../protocols/cohere-chat.js"
+import { OpenAIChat } from "../protocols/openai-chat.js"
+import { Route, type RouteDefaultsInput } from "../route/client.js"
+import { Endpoint } from "../route/endpoint.js"
+import { AuthOptions, type ProviderAuthOption } from "../route/auth-options.js"
+import { ProviderID, type ModelID, type OpenString } from "../schema/index.js"
+import type { ProviderPackage } from "../provider-package.js"
+
+export const id = ProviderID.make("cohere")
+const COMPATIBILITY_BASE_URL = "https://api.cohere.ai/compatibility/v1"
+export type ChatOptionsInput = { readonly reasoningEffort?: OpenString<"none" | "high"> }
+export type ProviderOptions = CohereChat.ProviderOptionsInput & ChatOptionsInput
+export type LanguageModelOptions = Omit<RouteDefaultsInput, "providerOptions"> &
+  ProviderAuthOption<"optional"> & {
+    readonly baseURL?: string
+    readonly providerOptions?: ProviderOptions
+  }
+export type Settings<Options = CohereChat.ProviderOptionsInput> = ProviderPackage.Settings &
+  Options & { readonly apiKey?: string; readonly baseURL?: string }
+
+export const route = CohereChat.route
+export const chatRoute = Route.make({
+  id: "cohere-chat-completions",
+  provider: id,
+  providerMetadataKey: "cohere",
+  protocol: OpenAIChat.protocol,
+  endpoint: Endpoint.path("/chat/completions", { baseURL: COMPATIBILITY_BASE_URL }),
+  framing: OpenAIChat.framing,
+})
+export const routes = [route, chatRoute]
+
+export const configure = (input: LanguageModelOptions = {}) => {
+  const { apiKey: _apiKey, auth: _auth, baseURL, ...defaults } = input
+  const auth = AuthOptions.bearer(input, "COHERE_API_KEY")
+  const native = route.with({ ...defaults, auth, endpoint: { baseURL: baseURL ?? CohereChat.DEFAULT_BASE_URL } })
+  const chat = chatRoute.with({ ...defaults, auth, endpoint: { baseURL: baseURL ?? COMPATIBILITY_BASE_URL } })
+  return {
+    id,
+    model: (modelID: string | ModelID) => native.model<CohereChat.ProviderOptionsInput>({ id: modelID }),
+    chat: (modelID: string | ModelID) =>
+      chat.model<ChatOptionsInput>({
+        id: modelID,
+        compatibility: {
+          maxTokensField: "max_tokens",
+          supportsStore: false,
+          supportsUsageInStreaming: true,
+          reasoningField: "reasoning_content",
+          supportsStrictMode: false,
+        },
+      }),
+    configure,
+  }
+}
+export const provider = configure()
+export const model: ProviderPackage.Definition<Settings, CohereChat.ProviderOptionsInput>["model"] = (
+  modelID,
+  { apiKey, baseURL, body, headers, ...providerOptions },
+) =>
+  configure({
+    apiKey,
+    baseURL,
+    headers,
+    http: body === undefined ? undefined : { body: { ...body } },
+    providerOptions,
+  }).model(modelID)
+export * as Cohere from "./cohere.js"
@@ -0,0 +1,16 @@
+import type { ProviderPackage } from "../../provider-package.js"
+import { Cohere } from "../cohere.js"
+
+export type Settings = Cohere.Settings<Cohere.ChatOptionsInput>
+
+export const model: ProviderPackage.Definition<Settings, Cohere.ChatOptionsInput>["model"] = (
+  modelID,
+  { apiKey, baseURL, body, headers, ...providerOptions },
+) =>
+  Cohere.configure({
+    apiKey,
+    baseURL,
+    headers,
+    http: body === undefined ? undefined : { body: { ...body } },
+    providerOptions,
+  }).chat(modelID)
@@ -0,0 +1,76 @@
+import type { ProviderPackage } from "../provider-package.js"
+import { OpenAIChat } from "../protocols/openai-chat.js"
+import { cacheControl } from "../protocols/utils/cache.js"
+import { AuthOptions, type ProviderAuthOption } from "../route/auth-options.js"
+import { Route, type RouteDefaultsInput } from "../route/client.js"
+import { Endpoint } from "../route/endpoint.js"
+import { Protocol } from "../route/protocol.js"
+import { ProviderID, type ModelID } from "../schema/index.js"
+import type { OpenAIProviderOptionsInput } from "./openai-options.js"
+
+export const id = ProviderID.make("digitalocean")
+const baseURL = "https://inference.do-ai.run/v1"
+
+export type LanguageModelOptions = Omit<RouteDefaultsInput, "providerOptions"> &
+  ProviderAuthOption<"optional"> & {
+    readonly baseURL?: string
+    readonly providerOptions?: OpenAIProviderOptionsInput
+  }
+
+export type Settings = ProviderPackage.Settings & OpenAIProviderOptionsInput & { readonly apiKey?: string }
+
+export const protocol = Protocol.make({
+  id: "digitalocean-chat",
+  body: {
+    schema: OpenAIChat.protocol.body.schema,
+    from: (request) => OpenAIChat.fromRequest(request, { cacheControl: cacheControl() }),
+  },
+  stream: OpenAIChat.protocol.stream,
+})
+
+export const route = Route.make({
+  id: "digitalocean",
+  provider: id,
+  providerMetadataKey: "digitalocean",
+  protocol,
+  endpoint: Endpoint.path("/chat/completions", { baseURL }),
+  framing: OpenAIChat.framing,
+})
+
+export const routes = [route]
+
+export const configure = (input: LanguageModelOptions = {}) => {
+  const { apiKey: _apiKey, auth: _auth, baseURL: endpoint, ...defaults } = input
+  const configured = route.with({
+    ...defaults,
+    endpoint: { baseURL: endpoint ?? baseURL },
+    auth: AuthOptions.bearer(input, ["DIGITALOCEAN_ACCESS_TOKEN", "DIGITALOCEAN_API_KEY", "DO_INFERENCE_API_KEY"]),
+  })
+  return {
+    id,
+    model: (modelID: string | ModelID) =>
+      configured.model<OpenAIProviderOptionsInput>({
+        id: modelID,
+        compatibility: {
+          supportsPromptCacheKey: true,
+        },
+      }),
+    configure,
+  }
+}
+
+export const provider = configure()
+
+export const model: ProviderPackage.Definition<Settings, OpenAIProviderOptionsInput>["model"] = (
+  modelID,
+  { apiKey, baseURL, body, headers, ...providerOptions },
+) =>
+  configure({
+    apiKey,
+    baseURL,
+    headers,
+    http: { body },
+    providerOptions,
+  }).model(modelID)
+
+export * as DigitalOcean from "./digitalocean.js"
@@ -3,8 +3,10 @@ import type { ProviderAuthOption } from "../route/auth-options.js"
 import { MediaRoute } from "../route/media.js"
 import { type HttpOptions, ProviderID, type ModelID } from "../schema/index.js"
 import { ElevenLabsSpeech } from "../protocols/elevenlabs-speech.js"
+import { ElevenLabsTranscription } from "../protocols/elevenlabs-transcription.js"

 export type { ElevenLabsOutputFormat, ElevenLabsSpeechOptions } from "../protocols/elevenlabs-speech.js"
+export type { ElevenLabsTranscriptionOptions } from "../protocols/elevenlabs-transcription.js"

 export const id = ProviderID.make("elevenlabs")

@@ -24,12 +26,15 @@ const auth = (options: ProviderAuthOption<"optional">) => {
 export const configure = (input: Config = {}) => {
  const media = MediaRoute.deployment(input, auth(input))
  const speech = (modelID: string | ModelID) => ElevenLabsSpeech.model({ ...media, id: modelID })
+  const transcription = (modelID: string | ModelID) => ElevenLabsTranscription.model({ ...media, id: modelID })
  return {
    id,
    speech,
+    transcription,
    configure,
  }
 }

 export const provider = configure()
 export const speech = provider.speech
+export const transcription = provider.transcription
@@ -37,7 +37,7 @@ export type Settings = ProviderPackage.Settings &
 const route = Route.make({
  id: "google-vertex-messages",
  provider: id,
-  providerMetadataKey: "anthropic",
+  providerMetadataKey: "vertex",
  protocol: Protocol.make({
    id: AnthropicMessages.protocol.id,
    body: {
@@ -71,7 +71,7 @@ export const protocol = Protocol.make({
 export const route = Route.make({
  id: "groq-chat",
  provider: id,
-  providerMetadataKey: "openai",
+  providerMetadataKey: "groq",
  protocol,
  endpoint: Endpoint.path("/chat/completions", { baseURL }),
  framing: OpenAIChat.framing,
@@ -9,11 +9,13 @@ export * as Baseten from "./baseten.js"
 export * as BlackForestLabs from "./black-forest-labs.js"
 export * as Cartesia from "./cartesia.js"
 export * as Cerebras from "./cerebras.js"
+export * as Cohere from "./cohere.js"
 export * as CloudflareAIGateway from "./cloudflare-ai-gateway.js"
 export * as CloudflareWorkersAI from "./cloudflare-workers-ai.js"
 export * as DeepInfra from "./deepinfra.js"
 export * as Deepgram from "./deepgram.js"
 export * as DeepSeek from "./deepseek.js"
+export * as DigitalOcean from "./digitalocean.js"
 export * as ElevenLabs from "./elevenlabs.js"
 export * as Fal from "./fal.js"
 export * as Fireworks from "./fireworks.js"
@@ -3,11 +3,11 @@ import { Route, type RouteDefaultsInput } from "../route/client.js"
 import { Endpoint } from "../route/endpoint.js"
 import { Protocol } from "../route/protocol.js"
 import { AuthOptions, type ProviderAuthOption } from "../route/auth-options.js"
-import { HttpOptions, ProviderID, type CacheHint, type ModelID, type OpenString } from "../schema/index.js"
+import { HttpOptions, ProviderID, type ModelID, type OpenString } from "../schema/index.js"
 import type { ProviderPackage } from "../provider-package.js"
 import { SystemOne } from "../experimental/system-one.js"
 import { OpenAIChat } from "../protocols/openai-chat.js"
-import { newBreakpoints, ttlBucket } from "../protocols/utils/cache.js"
+import { cacheControl } from "../protocols/utils/cache.js"
 import { isRecord, ProviderShared } from "../protocols/shared.js"

 export const id = ProviderID.make("openrouter")
@@ -131,18 +131,6 @@ export const protocol = Protocol.make({
  stream: OpenAIChat.protocol.stream,
 })

-const cacheControl = () => {
-  const breakpoints = newBreakpoints(4)
-  return (cache: CacheHint | undefined) => {
-    if (cache === undefined || breakpoints.remaining === 0) return undefined
-    breakpoints.remaining -= 1
-    return {
-      type: "ephemeral" as const,
-      ...(ttlBucket(cache.ttlSeconds) === "1h" ? { ttl: "1h" } : {}),
-    }
-  }
-}
-
 // OpenRouter forwards `reasoning.max_tokens` as the upstream thinking budget. Upstreams such as Anthropic and Alibaba
 // reject one that is not below the output limit; 1,024 is Anthropic's minimum budget.
 const fitReasoning = (reasoning: Record<string, unknown>, maxTokens: number | undefined) =>
@@ -7,8 +7,6 @@ import { OpenAIChat } from "../protocols/openai-chat.js"
 import { OpenResponsesChannel } from "../protocols/open-responses-channel.js"
 import { XAIResponses } from "../protocols/xai-responses.js"
 import { XAIImages } from "../protocols/xai-images.js"
-import { XAISpeech } from "../protocols/xai-speech.js"
-import { XAITranscription } from "../protocols/xai-transcription.js"
 import { XAIVideo } from "../protocols/xai-video.js"
 import type { OpenAIOptionsInput } from "./openai-options.js"
 import type { ProviderPackage } from "../provider-package.js"
@@ -31,21 +29,19 @@ export type Settings = ProviderPackage.Settings &
  }

 export type { XAIImageOptions } from "../protocols/xai-images.js"
-export type { XAISpeechOptions } from "../protocols/xai-speech.js"
-export type { XAITranscriptionOptions } from "../protocols/xai-transcription.js"
 export type { XAIVideoOptions } from "../protocols/xai-video.js"

 const RESPONSES_WEBSOCKET_ROTATE_AFTER_MS = 24 * 60 * 1000

 const responsesRoute = Route.make({
  compact: { endpoint: XAIResponses.compact },
-  id: "openai-responses",
+  id: "xai-responses",
  provider: id,
  providerMetadataKey: "xai",
  protocol: XAIResponses.protocol,
  endpoint: Endpoint.path("/responses", { baseURL }),
  transport: OpenResponsesChannel.transport({
-    id: "openai-responses",
+    id: "xai-responses",
    name: "xAI Responses",
    rotateAfterMs: RESPONSES_WEBSOCKET_ROTATE_AFTER_MS,
    // xAI continues a chain only from stored responses: with `store: false` (the route default) `previous_response_id`
@@ -57,7 +53,7 @@ const responsesRoute = Route.make({
 })

 const chatRoute = Route.make({
-  id: "openai-compatible-chat",
+  id: "xai-chat",
  provider: id,
  providerMetadataKey: "xai",
  protocol: OpenAIChat.protocol,
@@ -102,8 +98,6 @@ export const configure = (input: LanguageModelOptions = {}) => {
    chat,
    image: (modelID: string | ModelID) => XAIImages.model({ ...media, id: modelID }),
    video: (modelID: string | ModelID) => XAIVideo.model({ ...media, id: modelID }),
-    speech: (modelID: string | ModelID) => XAISpeech.model({ ...media, id: modelID }),
-    transcription: (modelID: string | ModelID) => XAITranscription.model({ ...media, id: modelID }),
    configure,
  }
 }
@@ -125,5 +119,3 @@ export const responses = provider.responses
 export const chat = provider.chat
 export const image = provider.image
 export const video = provider.video
-export const speech = provider.speech
-export const transcription = provider.transcription
@@ -8,7 +8,7 @@ import {
  HttpClientResponse,
 } from "effect/unstable/http"
 import { HttpContext, HttpRateLimitDetails, AIError, TransportError } from "../schema/index.js"
-import { classifyProviderFailure } from "../provider-error.js"
+import { classifyProviderFailure, providerErrorMessage } from "../provider-error.js"
 import { Service, type HttpMiddleware, type Interface } from "./executor-service.js"

 export { Service } from "./executor-service.js"
@@ -84,25 +84,17 @@ export const responseHttp = (response: HttpClientResponse.HttpClientResponse) =>
    headers: headerDetails(response.headers),
  })

-const decodeProviderBody = Schema.decodeUnknownOption(
-  Schema.fromJsonString(
-    Schema.Struct({
-      message: Schema.optionalKey(Schema.String),
-      // xAI sends `{ code, error }` with the readable reason as a plain string.
-      error: Schema.optionalKey(
-        Schema.Union([Schema.String, Schema.Struct({ message: Schema.optionalKey(Schema.String) })]),
-      ),
-    }),
-  ),
-)
+const MAX_BODY_CHARS = 2000

+// Without a recognized message, show the raw body so the provider's explanation is never dropped.
 const providerMessage = (status: number, body: string | void) => {
-  const decoded = body === undefined ? undefined : Option.getOrUndefined(decodeProviderBody(body))
-  const error = typeof decoded?.error === "string" ? decoded.error : decoded?.error?.message
-  return (
-    [error, decoded?.message].find((message) => message?.trim()) ??
-    `Provider request failed with HTTP ${status}`
-  )
+  const fallback = `Provider request failed with HTTP ${status}`
+  const text = body?.trim() ?? ""
+  const message = providerErrorMessage(text)
+  if (message) return message
+  // Gateway and proxy HTML error pages are markup, not an explanation.
+  if (!text || /^<(?:!doctype|html)/i.test(text)) return fallback
+  return `${fallback}: ${text.length > MAX_BODY_CHARS ? `${text.slice(0, MAX_BODY_CHARS)}…` : text}`
 }

 const statusError = (response: HttpClientResponse.HttpClientResponse) =>
@@ -167,6 +159,15 @@ const nativeTransportFailure = (error: unknown) => {
  return failure
 }

+// HTTP hooks re-wrap response bodies, so a read failure can arrive as an HttpClientError caused by
+// another HttpClientError. The innermost cause carries the native failure.
+const rootCause = (error: unknown): unknown =>
+  HttpClientError.isHttpClientError(error) && "cause" in error.reason && error.reason.cause !== undefined
+    ? rootCause(error.reason.cause)
+    : error
+
+const CONNECTION_LOST = "Connection lost while reading the response"
+
 const httpError = (input: {
  readonly error: unknown
  readonly request: HttpClientRequest.HttpClientRequest
@@ -187,20 +188,21 @@ const httpError = (input: {
      }),
    })

-  const source =
-    HttpClientError.isHttpClientError(input.error) && "cause" in input.error.reason
-      ? (input.error.reason.cause ?? input.error)
-      : input.error
+  const source = rootCause(input.error)
  const native = nativeTransportFailure(source)
  const code = native?.code
-  const raw = native?.message ?? (input.error instanceof Error ? input.error.message : undefined)
-  const detail = raw
-  const message = code && detail && !detail.includes(code) ? `${code}: ${detail}` : detail
+  const detail =
+    code && native?.message && !native.message.includes(code) ? `${code}: ${native.message}` : native?.message
+  const message = detail ?? (input.error instanceof Error ? input.error.message : undefined)

  if (Cause.isTimeoutError(input.error) || Cause.isTimeoutError(source))
    return transportError({ message: message ?? "HTTP transport timed out", code: code ?? "Timeout" })
  if (!HttpClientError.isHttpClientError(input.error))
    return transportError({ message: message ?? "HTTP transport failed", code })
+  // Effect reports every response body read failure as a DecodeError, but the raw byte stream decodes
+  // nothing: provider output parsing happens later and fails as InvalidProviderOutput.
+  if (input.operation === "read" && input.error.reason._tag === "DecodeError")
+    return transportError({ message: detail ? `${CONNECTION_LOST}: ${detail}` : CONNECTION_LOST, code })
  if (input.error.reason._tag === "TransportError") {
    return transportError({
      message: message ?? input.error.reason.description ?? "HTTP transport failed",
@@ -5,12 +5,14 @@ import { Media } from "../media.js"
 import type { AuthInput } from "./auth.js"
 import {
  AIError,
+  AuthenticationError,
  ContentPolicyError,
  HttpContext,
  InvalidProviderOutputError,
  InvalidRequestError,
  ProviderID,
  ProviderInternalError,
+  RateLimitError,
  UnsupportedOperationError,
 } from "../schema/index.js"

@@ -137,6 +139,11 @@ export interface Queued<Request, Response, Token> {
  readonly cancel?: {
    readonly method: AuthInput["method"]
    readonly path: (token: Token) => string
+    /**
+     * Fetch a fresh status first and skip the call for terminal generations, for providers whose cancel endpoint
+     * destroys finished work (Runway's `DELETE /v1/tasks/{id}` deletes completed tasks and their outputs).
+     */
+    readonly activeOnly?: boolean
  }
 }

@@ -183,6 +190,16 @@ export const stream = <Request, Event, Frame, State>(
 // Response helpers
 // ---------------------------------------------------------------------------

+/** Reasons a provider can report for a `failed` generation; anything it does not classify is `ProviderInternal`. */
+const FAILURES = {
+  InvalidRequest: InvalidRequestError,
+  Authentication: AuthenticationError,
+  RateLimit: RateLimitError,
+  ProviderInternal: ProviderInternalError,
+}
+
+export type Failure = keyof typeof FAILURES
+
 const context = (response: HttpClientResponse.HttpClientResponse) =>
  new HttpContext({ url: response.request.url, status: response.status, headers: response.headers })

@@ -194,8 +211,10 @@ export const identity = (input: { readonly id: string; readonly name: string; re

  /**
   * Read a text body while retaining the original payload and HTTP context on every downstream error. `invalid` is a
-   * malformed provider document; `ended` is a generation that reached a terminal status without output (`failed` is
-   * provider-side, `cancelled`/`expired` mean the result will never exist); `contentPolicy` is a moderated result.
+   * malformed provider document; `ended` is a generation that reached a terminal status without output (`failed`
+   * carries the provider's classification, defaulting to `ProviderInternal`; `cancelled`/`expired` mean the result
+   * will never exist); `pending` is a `result()` read before the generation finished, which is caller misuse;
+   * `contentPolicy` is a moderated result.
   */
  const text = Effect.fn("MediaProtocol.text")(function* (response: HttpClientResponse.HttpClientResponse) {
    const http = context(response)
@@ -217,13 +236,25 @@ export const identity = (input: { readonly id: string; readonly name: string; re
      http,
      invalid: (message: string, cause?: unknown) =>
        new AIError({ reason: new InvalidProviderOutputError({ route: input.id, message, body, http, cause }) }),
-      ended: (status: Exclude<Status, "queued" | "running" | "completed">, message: string) =>
+      ended: (
+        status: Exclude<Status, "queued" | "running" | "completed">,
+        message: string,
+        failure: Failure = "ProviderInternal",
+      ) =>
        new AIError({
          reason:
            status === "failed"
-              ? new ProviderInternalError({ message, body, http })
+              ? new FAILURES[failure]({ message, body, http })
              : new InvalidRequestError({ message, body, http }),
        }),
+      pending: (id: string) =>
+        new AIError({
+          reason: new InvalidRequestError({
+            message: `${input.name} generation ${id} has not finished; await it before reading the result`,
+            body,
+            http,
+          }),
+        }),
      contentPolicy: (message: string) => new AIError({ reason: new ContentPolicyError({ message, body, http }) }),
    }
  })
@@ -285,11 +316,14 @@ export const status = <Table extends Record<string, Status>>(
  raw: string,
  output: Output,
 ): Effect.Effect<Status, AIError> => {
-  const normalized: Status | undefined = table[raw]
-  if (normalized === undefined) return Effect.fail(output.invalid(`Unknown generation status "${raw}"`))
-  return Effect.succeed(normalized)
+  if (!Object.hasOwn(table, raw)) return Effect.fail(output.invalid(`Unknown generation status "${raw}"`))
+  return Effect.succeed(table[raw])
 }

+/** Map a provider error code through the protocol's table; missing or unmapped codes are `ProviderInternal`. */
+export const failure = (table: Readonly<Record<string, Failure>>, code: string | number | undefined): Failure =>
+  code !== undefined && Object.hasOwn(table, code) ? table[code] : "ProviderInternal"
+
 /** A `url` asset whose provider-declared retention window starts now. */
 export const expiringUrl = (url: string, retention: Duration.Duration, options?: Parameters<typeof Media.url>[1]) =>
  Clock.currentTimeMillis.pipe(
@@ -1,12 +1,13 @@
-import { Effect, Schema, Stream } from "effect"
+import { Duration, Effect, Schedule, Schema, Stream } from "effect"
 import { Headers, HttpClientRequest, type HttpClientResponse } from "effect/unstable/http"
 import { Auth, type AuthInput } from "./auth.js"
 import { Endpoint } from "./endpoint.js"
 import { RequestExecutorService, type Interface } from "./executor-service.js"
 import { RequestExecutor } from "./executor.js"
 import { MediaProtocol } from "./media-protocol.js"
-import { Generation } from "../generation.js"
+import { Generation, isTerminal } from "../generation.js"
 import type { Media } from "../media.js"
+import { isRetryable } from "../provider-error.js"
 import {
  AIError,
  AIErrorReason,
@@ -137,6 +138,32 @@ export const inline = <Request extends MediaRequest, Response>(
  }
 }

+const READ_RETRY_MAX_DELAY = Duration.seconds(30)
+
+/**
+ * Status and result reads retry transient failures; `start` and `cancel` never do. Gaps grow exponentially from 1s,
+ * jittered, up to 30s each, for at most 8 retries (about two minutes when every attempt fails), so a direct
+ * `Generation.result()` stays bounded; `await` and `events` also cut retries off at `poll.timeout`. A provider
+ * `retryAfterMs` raises the gap, still capped at 30s.
+ */
+const READ_RETRY = Schedule.max([
+  Schedule.min([Schedule.exponential("1 second"), Schedule.spaced(READ_RETRY_MAX_DELAY)]),
+  Schedule.recurs(8),
+]).pipe(
+  Schedule.jittered,
+  Schedule.setInputType<AIError>(),
+  Schedule.modifyDelay(({ input, duration }) =>
+    Effect.succeed(
+      Duration.min(
+        input.reason._tag === "RateLimit" || input.reason._tag === "ProviderInternal"
+          ? Duration.max(duration, Duration.millis(input.reason.retryAfterMs ?? 0))
+          : duration,
+        READ_RETRY_MAX_DELAY,
+      ),
+    ),
+  ),
+)
+
 /**
 * Compose a queued media protocol the same way, adding `start`/`resume` handles whose polls reuse the route's auth,
 * deployment headers, and (for `start`) the request's `http` overlay. The token is decoded once at the boundary and
@@ -154,6 +181,8 @@ export const queued = <Request extends MediaRequest, Response, Token>(
  const generationRoute = (token: Token, http: HttpOptions | undefined, execute: Execute) => {
    const materialize = (asset: Media.Asset) =>
      asset.materialize().pipe(Effect.provideService(RequestExecutorService, { execute }))
+    // Only the GET exchange retries: a decoded terminal failure (`output.ended`) can be a `ProviderInternal` too, and
+    // re-reading it would spin until the caller's deadline.
    const poll = <A>(operation: {
      readonly path: (token: Token) => string
      readonly decode: (
@@ -161,17 +190,23 @@ export const queued = <Request extends MediaRequest, Response, Token>(
        context: MediaProtocol.PollContext<Token>,
      ) => Effect.Effect<A, AIError>
    }) =>
-      transport
-        .call("GET", operation.path(token), http, execute)
-        .pipe(Effect.flatMap((sent) => operation.decode(sent.response, { token, auth: sent.auth, materialize })))
+      transport.call("GET", operation.path(token), http, execute).pipe(
+        Effect.retry({ schedule: READ_RETRY, while: isRetryable }),
+        Effect.flatMap((sent) => operation.decode(sent.response, { token, auth: sent.auth, materialize })),
+      )
+    const status = poll(protocol.status)
    const cancel = protocol.cancel
+    const send =
+      cancel === undefined
+        ? undefined
+        : transport.call(cancel.method, cancel.path(token), http, execute).pipe(Effect.asVoid)
    return {
-      status: poll(protocol.status),
+      status,
      result: poll(protocol.result),
      cancel:
-        cancel === undefined
-          ? undefined
-          : transport.call(cancel.method, cancel.path(token), http, execute).pipe(Effect.asVoid),
+        send !== undefined && cancel?.activeOnly
+          ? status.pipe(Effect.flatMap((snapshot) => (isTerminal(snapshot.status) ? Effect.void : send)))
+          : send,
    }
  }

@@ -380,7 +415,11 @@ const encode = (body: MediaProtocol.Body | undefined, headers: Headers.Headers)
  }
 }

-/** Common fields are never silently dropped: a present field the protocol declared unsupported fails typed. */
+/**
+ * Common fields are never silently dropped: a present field the protocol declared unsupported fails typed. `false`
+ * counts as present because some booleans mean something when false (video `audio`); protocols reject opt-in
+ * booleans such as speech `timestamps` with `=== true` in `body.from` instead of listing them.
+ */
 const rejectUnsupported = <Request extends object>(
  route: string,
  provider: ProviderID,
@@ -1,11 +1,19 @@
-import { Effect } from "effect"
+import { Clock, Duration, Effect, Stream } from "effect"
 import { Headers, HttpClientRequest } from "effect/unstable/http"
 import { Auth } from "../auth.js"
 import { render as renderEndpoint } from "../endpoint.js"
 import { Framing } from "../framing.js"
 import type { HttpMiddleware, Transport, TransportPrepareInput } from "./index.js"
 import * as ProviderShared from "../../protocols/shared.js"
-import { mergeJsonRecords, type LLMRequest } from "../../schema/index.js"
+import {
+  AIError,
+  DEFAULT_HTTP_TIMEOUT_MS,
+  mergeJsonRecords,
+  TransportError,
+  type HttpContext,
+  type HttpTimeout,
+  type LLMRequest,
+} from "../../schema/index.js"
 import { RequestExecutor } from "../executor.js"

 export type JsonRequestInput<Body> = TransportPrepareInput<Body>
@@ -87,17 +95,53 @@ export const httpJson = <Body, Frame>(input: HttpJsonInput<Body, Frame>): HttpJs
        middleware: prepareInput.middleware,
      }
    }),
-  execute: (prepared, _request, runtime) =>
+  execute: (prepared, request, runtime) =>
    Effect.gen(function* () {
-      const response = yield* runtime.http.execute(prepared.request, prepared.middleware)
+      const timeout = (operation: "request" | "read", message: string, http?: HttpContext) =>
+        new AIError({
+          reason: new TransportError({
+            message,
+            transport: "http",
+            operation,
+            code: "Timeout",
+            url: prepared.request.url,
+            http,
+          }),
+        })
+      const started = yield* Clock.currentTimeMillis
+      // Unlike the header and chunk limits, the whole-request budget has no default.
+      const total = request.http?.timeout ? Duration.millis(request.http.timeout) : Duration.infinity
+      const response = yield* runtime.http.execute(prepared.request, prepared.middleware).pipe(
+        Effect.timeoutOrElse({
+          duration: Duration.min(timeoutDuration(request.http?.headerTimeout), total),
+          orElse: () => timeout("request", "Timed out waiting for response headers"),
+        }),
+      )
+      const http = RequestExecutor.responseHttp(response)
+      const remaining = Duration.subtract(total, Duration.millis((yield* Clock.currentTimeMillis) - started))
      return {
-        frames: prepared.framing.frame(RequestExecutor.responseStream(response)),
-        http: RequestExecutor.responseHttp(response),
+        frames: prepared.framing.frame(
+          RequestExecutor.responseStream(response).pipe(
+            Stream.timeoutOrElse({
+              duration: timeoutDuration(request.http?.chunkTimeout),
+              orElse: () => Stream.fail(timeout("read", "Timed out waiting for response data", http)),
+            }),
+            Stream.interruptWhen(
+              Effect.sleep(remaining).pipe(
+                Effect.andThen(Effect.fail(timeout("read", "Timed out waiting for the response to complete", http))),
+              ),
+            ),
+          ),
+        ),
+        http,
        body: prepared.framing.body,
      }
    }),
 })

+const timeoutDuration = (value: HttpTimeout | undefined) =>
+  value === false ? Duration.infinity : Duration.millis(value ?? DEFAULT_HTTP_TIMEOUT_MS)
+
 export const sseJson = {
  id: "http-json/sse",
  with: <Body>() => httpJson<Body, string>({ framing: Framing.sse }),
@@ -264,6 +264,10 @@ export type ToolError = Schema.Schema.Type<typeof ToolError>
 export const FinishReasonDetails = Schema.Struct({
  normalized: FinishReason,
  raw: Schema.optional(Schema.String),
+  /** The provider's policy area for a content-filter finish, such as `cyber`. */
+  category: Schema.optional(Schema.String),
+  /** The provider's human-readable reason for a content-filter finish. */
+  explanation: Schema.optional(Schema.String),
 }).annotate({ identifier: "LLM.FinishReasonDetails" })
 export type FinishReasonDetails = Schema.Schema.Type<typeof FinishReasonDetails>

@@ -48,10 +48,23 @@ export const mergeProviderOptions = (
  ...items: ReadonlyArray<ProviderOptions | undefined>
 ): ProviderOptions | undefined => mergeJsonRecords(...items)

+/** Milliseconds for an HTTP timeout, or `false` to disable it. */
+export const HttpTimeout = Schema.Union([Schema.Number.check(Schema.isGreaterThan(0)), Schema.Literal(false)])
+export type HttpTimeout = Schema.Schema.Type<typeof HttpTimeout>
+
+/** Applied to `headerTimeout` and `chunkTimeout` when a request leaves them unset. */
+export const DEFAULT_HTTP_TIMEOUT_MS = 300_000
+
 export class HttpOptions extends Schema.Class<HttpOptions>("AI.HttpOptions")({
  body: Schema.optional(JsonSchema),
  headers: Schema.optional(Schema.Record(Schema.String, Schema.String)),
  query: Schema.optional(Schema.Record(Schema.String, Schema.String)),
+  /** Time allowed for the whole request, from send until the response completes. Unbounded when unset. */
+  timeout: Schema.optional(HttpTimeout),
+  /** Time allowed for response headers to arrive. */
+  headerTimeout: Schema.optional(HttpTimeout),
+  /** Time allowed between streamed response chunks once headers have arrived. */
+  chunkTimeout: Schema.optional(HttpTimeout),
 }) {}

 export namespace HttpOptions {
@@ -70,8 +83,12 @@ export const mergeHttpOptions = (...items: ReadonlyArray<HttpOptions | undefined
  const body = mergeJsonRecords(...items.map((item) => item?.body))
  const headers = mergeStringRecords(...items.map((item) => item?.headers))
  const query = mergeStringRecords(...items.map((item) => item?.query))
-  if (!body && !headers && !query) return undefined
-  return new HttpOptions({ body, headers, query })
+  const timeout = items.findLast((item) => item?.timeout !== undefined)?.timeout
+  const headerTimeout = items.findLast((item) => item?.headerTimeout !== undefined)?.headerTimeout
+  const chunkTimeout = items.findLast((item) => item?.chunkTimeout !== undefined)?.chunkTimeout
+  if (!body && !headers && !query && timeout === undefined && headerTimeout === undefined && chunkTimeout === undefined)
+    return undefined
+  return new HttpOptions({ body, headers, query, timeout, headerTimeout, chunkTimeout })
 }

 export class GenerationOptions extends Schema.Class<GenerationOptions>("LLM.GenerationOptions")({
@@ -6,10 +6,17 @@ const MISSING_TOOL_RESULT = "Tool result missing"
 export function normalizeToolHistory(messages: ReadonlyArray<Message>) {
  const normalized: Message[] = []
  const pending = new Map<string, ToolCallPart>()
+  // System updates cannot sit between a tool call and its results, so they wait until every pending call is answered.
+  const held: Message[] = []
+  const releaseHeld = () => {
+    if (pending.size > 0) return
+    normalized.push(...held)
+    held.length = 0
+  }
  const appendMissingResults = () => {
-    if (pending.size === 0) return
-    normalized.push(missingToolResults(pending.values()))
+    if (pending.size > 0) normalized.push(missingToolResults(pending.values()))
    pending.clear()
+    releaseHeld()
  }

  for (const message of messages) {
@@ -18,6 +25,12 @@ export function normalizeToolHistory(messages: ReadonlyArray<Message>) {
    if (message.role === "tool") {
      const tool = normalizeToolMessage(message, pending)
      if (tool) normalized.push(tool)
+      releaseHeld()
+      continue
+    }
+
+    if (message.role === "system" && pending.size > 0) {
+      held.push(message)
      continue
    }

@@ -27,6 +40,7 @@ export function normalizeToolHistory(messages: ReadonlyArray<Message>) {
      if (part.type === "tool-call" && part.providerExecuted !== true) pending.set(part.id, part)
    }
  }
+  appendMissingResults()

  return normalized.length === messages.length && normalized.every((message, index) => message === messages[index])
    ? messages
@@ -103,7 +103,7 @@ export type TranscriptionRequestInput<Model extends TranscriptionModel = Transcr
 // Response and events
 // ---------------------------------------------------------------------------

-/** Speaker labels are provider-native (`A`, `0`, `spk:0`, or a known speaker name). */
+/** Speaker labels are provider-native (`A`, `0`, `spk:0`, `speaker_0`, or a known speaker name). */
 export const TranscriptionSegment = Schema.Struct({
  text: Schema.String,
  startSeconds: Schema.Number,
@@ -3,7 +3,18 @@ import { Effect } from "effect"
 import { CacheHint, LLM, Message } from "../src/index.js"
 import { Auth } from "../src/route.js"
 import { compileRequest } from "../src/route/client.js"
-import { AmazonBedrock, GoogleVertexMessages } from "../src/providers.js"
+import {
+  Alibaba,
+  AmazonBedrock,
+  AnthropicCompatible,
+  CloudflareAIGateway,
+  DigitalOcean,
+  GoogleVertexMessages,
+  Meta,
+  MiniMax,
+  Moonshot,
+  ZAICodingPlan,
+} from "../src/providers.js"
 import * as AnthropicMessages from "../src/protocols/anthropic-messages.js"
 import * as Gemini from "../src/protocols/gemini.js"
 import * as OpenAIChat from "../src/protocols/openai-chat.js"
@@ -107,6 +118,134 @@ describe("applyCachePolicy", () => {
    }),
  )

+  it.effect("'auto' emits cache_control markers on DigitalOcean", () =>
+    Effect.gen(function* () {
+      const prepared = yield* compileRequest(
+        LLM.request({
+          model: DigitalOcean.configure({ apiKey: "test" }).model("anthropic-claude-fable-5.1"),
+          system: "You are concise.",
+          tools: [{ name: "lookup", description: "Look up a value", inputSchema: { type: "object", properties: {} } }],
+          prompt: "hi",
+        }),
+      )
+
+      expect(prepared.body).toMatchObject({
+        tools: [{ type: "function", function: { name: "lookup" }, cache_control: { type: "ephemeral" } }],
+        messages: [
+          {
+            role: "system",
+            content: [{ text: "You are concise.", cache_control: { type: "ephemeral" } }],
+          },
+          {
+            role: "user",
+            content: [{ text: "hi", cache_control: { type: "ephemeral" } }],
+          },
+        ],
+      })
+    }),
+  )
+
+  it.effect("Alibaba chat caches the system and conversation tail without marking tools", () =>
+    Effect.gen(function* () {
+      const prepared = yield* compileRequest(
+        LLM.request({
+          model: Alibaba.configure({ region: "ap-southeast-1", apiKey: "test" }).chat("qwen3.8-max"),
+          system: "You are concise.",
+          tools: [{ name: "lookup", description: "Look up a value", inputSchema: { type: "object", properties: {} } }],
+          prompt: "hi",
+        }),
+      )
+
+      expect(prepared.body).toMatchObject({
+        tools: [{ type: "function", function: { name: "lookup" } }],
+        messages: [
+          {
+            role: "system",
+            content: [{ text: "You are concise.", cache_control: { type: "ephemeral" } }],
+          },
+          { role: "user", content: [{ text: "hi", cache_control: { type: "ephemeral" } }] },
+        ],
+      })
+      expect(prepared.body.tools?.[0]?.cache_control).toBeUndefined()
+    }),
+  )
+
+  it.effect("Alibaba chat omits automatic cache markers when cache is none", () =>
+    Effect.gen(function* () {
+      const prepared = yield* compileRequest(
+        LLM.request({
+          model: Alibaba.configure({ region: "ap-southeast-1", apiKey: "test" }).chat("qwen3.8-max"),
+          system: "You are concise.",
+          prompt: "hi",
+          cache: "none",
+        }),
+      )
+
+      expect(JSON.stringify(prepared.body)).not.toContain("cache_control")
+    }),
+  )
+
+  it.effect("Alibaba chat does not assume non-Qwen models support cache markers", () =>
+    Effect.gen(function* () {
+      const alibaba = Alibaba.configure({ region: "ap-southeast-1", apiKey: "test" })
+      for (const modelID of ["kimi-k3", "glm-5.2", "deepseek-v4-flash-0731", "MiniMax-M2.5"]) {
+        const prepared = yield* compileRequest(
+          LLM.request({ model: alibaba.chat(modelID), system: "You are concise.", prompt: "hi" }),
+        )
+
+        expect(JSON.stringify(prepared.body)).not.toContain("cache_control")
+      }
+    }),
+  )
+
+  it.effect("'auto' emits Anthropic cache markers on Anthropic-compatible routes", () =>
+    Effect.gen(function* () {
+      const prepared = yield* compileRequest(
+        LLM.request({
+          model: AnthropicCompatible.configure({ apiKey: "test", baseURL: "https://messages.example.test/v1" }).model(
+            "compatible",
+          ),
+          system: "You are concise.",
+          prompt: "hi",
+        }),
+      )
+
+      expect(prepared.route).toBe("anthropic-compatible-messages")
+      expect(prepared.body).toMatchObject({
+        system: [{ type: "text", text: "You are concise.", cache_control: { type: "ephemeral" } }],
+        messages: [{ role: "user", content: [{ type: "text", text: "hi", cache_control: { type: "ephemeral" } }] }],
+      })
+    }),
+  )
+
+  const messagesModels = [
+    ["alibaba-messages", Alibaba.configure({ region: "ap-southeast-1", apiKey: "test" }).messages("qwen3.8-max")],
+    [
+      "cloudflare-ai-gateway-messages",
+      CloudflareAIGateway.configure({ accountId: "test", gatewayId: "test", apiKey: "test" }).model(
+        "anthropic/claude-sonnet-4-6",
+      ),
+    ],
+    ["meta-messages", Meta.configure({ apiKey: "test" }).messages("muse-spark-1.3")],
+    ["minimax-messages", MiniMax.configure({ apiKey: "test" }).model("MiniMax-M3")],
+    ["moonshot-messages", Moonshot.configure({ apiKey: "test" }).messages("kimi-k3")],
+    ["zai-coding-messages", ZAICodingPlan.configure({ apiKey: "test" }).messages("glm-5.3")],
+  ] as const
+
+  messagesModels.forEach(([route, model]) =>
+    it.effect(`'auto' emits cache markers on ${route}`, () =>
+      Effect.gen(function* () {
+        const prepared = yield* compileRequest(LLM.request({ model, system: "Sys", prompt: "hi" }))
+
+        expect(prepared.route).toBe(route)
+        expect(prepared.body).toMatchObject({
+          system: [{ type: "text", text: "Sys", cache_control: { type: "ephemeral" } }],
+          messages: [{ role: "user", content: [{ type: "text", text: "hi", cache_control: { type: "ephemeral" } }] }],
+        })
+      }),
+    ),
+  )
+
  it.effect("'auto' is a no-op on OpenAI (implicit caching protocol)", () =>
    Effect.gen(function* () {
      const prepared = yield* compileRequest(
@@ -147,7 +147,7 @@ describe("Anthropic Messages effort updates", () => {
    }),
  )

-  it.effect("accepts a marker between a tool call and its result", () =>
+  it.effect("moves a marker between a tool call and its result after the result", () =>
    Effect.gen(function* () {
      const prepared = yield* compileRequest(
        LLM.request({
@@ -166,8 +166,8 @@ describe("Anthropic Messages effort updates", () => {
      expect(prepared.body.messages).toEqual([
        { role: "user", content: [{ type: "text", text: "Weather?" }] },
        { role: "assistant", content: [{ type: "tool_use", id: "call_1", name: "lookup", input: {} }] },
-        { role: "system", content: [], output_config: { effort: "low" } },
        { role: "user", content: [{ type: "tool_result", tool_use_id: "call_1", content: '{"temp":72}' }] },
+        { role: "system", content: [], output_config: { effort: "low" } },
      ])
    }),
  )
@@ -197,9 +197,15 @@ describe("Anthropic Messages effort updates", () => {
    ["anthropic/claude-opus-5", true],
    ["claude-fable-5-1", true],
    ["claude-mythos-5-1", true],
+    ["claude-opus-5-5", true],
+    ["claude-sonnet-5-5", true],
+    ["anthropic/claude-sonnet-5-5", true],
+    ["claude-sonnet-6", true],
+    ["claude-haiku-6", true],
    ["claude-fable-5", false],
    ["claude-opus-4-8", false],
    ["claude-sonnet-5", false],
+    ["claude-sonnet-5-20260801", false],
    ["kimi-k2.5", false],
  ] as const) {
    it.effect(`${supported ? "lowers" : "strips"} markers for ${id}`, () =>
@@ -351,11 +357,48 @@ describe("OpenAI Responses effort updates", () => {
    }),
  )

+  it.effect("strips markers when the body overlay selects pro reasoning mode", () =>
+    Effect.gen(function* () {
+      const prepared = yield* compileRequest(
+        LLM.request({
+          model: OpenAI.configure({ apiKey: "fixture", http: { body: { reasoning: { mode: "pro" } } } }).responses(
+            "gpt-6-sol",
+          ),
+          messages: conversation,
+          providerOptions: { reasoningEffort: "low" },
+        }),
+      )
+
+      expect(updates(prepared.body)).toEqual([])
+      expect(prepared.body.reasoning).toEqual({ effort: "low" })
+    }),
+  )
+
  for (const [id, supported] of [
    ["gpt-6-astra", true],
    ["openai/gpt-6-astra", true],
-    ["gpt-6-astra-2026-09-01", false],
+    ["gpt-6-sol", true],
+    ["openai/gpt-6-sol", true],
+    ["gpt-6-luna", true],
+    ["openai/gpt-6-luna", true],
+    ["gpt-6", true],
+    ["gpt-6.1-sol", true],
+    ["openai/gpt-6.1-sol", true],
+    ["GPT-6.1-SOL", true],
+    ["gpt-6-astra-2026-09-01", true],
+    ["gpt-6-sol-pro", true],
+    ["gpt-6-luna-pro", true],
+    ["gpt-6-sol-fast", true],
+    ["gpt-7", true],
+    ["openai/gpt-7.2-new-family", true],
+    ["gpt-10.1", true],
+    ["gpt-5", false],
    ["gpt-5.6-sol", false],
+    ["gpt-5.10", false],
+    ["not-gpt-6-sol", false],
+    ["gpt-6foo", false],
+    ["gpt-6.x-sol", false],
+    ["future-model", false],
  ] as const) {
    it.effect(`${supported ? "lowers" : "strips"} markers for ${id}`, () =>
      Effect.gen(function* () {
@@ -39,17 +39,18 @@ describe("experimental Evaluation", () => {
        type: "choice",
        choice: "billing",
        probabilities: { billing: 0.9, technical: 0.1 },
+        confidence: 0.8,
      })
      expect(response.answers.urgency).toEqual({
        type: "score",
        score: 1.2,
        probabilities: { "0": 0, "1": 0.8, "2": 0.2 },
+        confidence: 0.6,
      })
      expect(response.answers.refund).toEqual({ type: "boolean", probability: 0.97 })
      expect(response.usage?.totalTokens).toBe(36)
      expect(response.providerMetadata).toEqual({
        typesafe: {
-          confidence: { department: 0.8, urgency: 0.6 },
          legend: { urgency: { "0": "Can wait", "1": "Needs attention", "2": "Blocking" } },
        },
      })
@@ -26,8 +26,10 @@ const request = Evaluation.request({
 const result = EvaluationClient.evaluate(request)
 type Result = Success<typeof result>
 type Choice = Assert<Equal<Result["answers"]["topic"]["choice"], "billing" | "support">>
+type Confidence = Assert<Equal<Result["answers"]["topic"]["confidence"], number | undefined>>
 type ClientRequirements = Assert<Equal<Requirements<typeof result>, Service>>
 void (true satisfies Choice)
+void (true satisfies Confidence)
 void (true satisfies ClientRequirements)

 Effect.gen(function* () {
@@ -1,11 +1,17 @@
 import { describe, expect } from "bun:test"
 import { Deferred, Effect, Fiber, Layer, Ref, Stream } from "effect"
-import { Headers, HttpClientError, HttpClientRequest } from "effect/unstable/http"
-import { LLM, AIError, HttpContext, InvalidProviderOutputError, TransportError } from "../src/index.js"
-import { LLMClient, RequestExecutor, WebSocketTransport, type WebSocketChannelExecutor } from "../src/route.js"
+import { Headers, HttpClientError, HttpClientRequest, HttpClientResponse } from "effect/unstable/http"
+import { LLM, AIError, HttpContext, InvalidProviderOutputError, TransportError, isRetryable } from "../src/index.js"
+import {
+  LLMClient,
+  RequestExecutor,
+  WebSocketTransport,
+  type HttpMiddleware,
+  type WebSocketChannelExecutor,
+} from "../src/route.js"
 import { route } from "../src/protocols/openai-chat.js"
 import { configure } from "../src/providers/openai.js"
-import { dynamicResponse, fixedResponse, handlerLayer, systemError } from "./lib/http.js"
+import { dynamicResponse, fixedResponse, handlerLayer, systemError, truncatedStream } from "./lib/http.js"
 import { deltaChunk } from "./lib/openai-chunks.js"
 import { sseEvents, sseRaw } from "./lib/sse.js"
 import { it } from "./lib/effect.js"
@@ -18,6 +24,8 @@ const secretRequest = HttpClientRequest.post("https://provider.test/v1/chat?api_
  HttpClientRequest.setHeaders(Headers.fromInput({ authorization: "Bearer header-secret-456" })),
 )

+const sseRequest = HttpClientRequest.post("https://provider.test/v1/messages")
+
 const expectAIError = (error: unknown) => {
  expect(error).toBeInstanceOf(AIError)
  if (!(error instanceof AIError)) throw new Error("expected AIError")
@@ -93,7 +101,9 @@ describe("RequestExecutor", () => {
      const error = yield* RequestExecutor.stream(executor, secretRequest).pipe(Stream.runDrain, Effect.flip)

      expectAIError(error)
-      expect(error.message).toBe("ECONNRESET: disconnected query-secret-123 header-secret-456")
+      expect(error.message).toBe(
+        "Connection lost while reading the response: ECONNRESET: disconnected query-secret-123 header-secret-456",
+      )
      expect(error.reason.http).toMatchObject({ status: 200, url: secretRequest.url })
      expect(error.reason.cause).toMatchObject({ code: "ECONNRESET" })
      expect(error.reason).toMatchObject({
@@ -123,7 +133,7 @@ describe("RequestExecutor", () => {
      const error = yield* RequestExecutor.stream(executor, secretRequest).pipe(Stream.runDrain, Effect.flip)

      expectAIError(error)
-      expect(error.message).toBe("ECONNRESET: socket closed")
+      expect(error.message).toBe("Connection lost while reading the response: ECONNRESET: socket closed")
      expect(error.reason.cause).toBeInstanceOf(TypeError)
      expect(error.reason).toMatchObject({
        _tag: "Transport",
@@ -144,6 +154,54 @@ describe("RequestExecutor", () => {
    ),
  )

+  it.effect("reports a connection lost mid-stream through middleware that re-wraps the body", () =>
+    Effect.gen(function* () {
+      const executor = yield* RequestExecutor.Service
+      const chunks: Array<Uint8Array> = []
+      // Session HTTP hooks hand plugins a web Response, so the body is re-wrapped around the original stream.
+      const rewrap: HttpMiddleware = (input, handler) =>
+        Effect.gen(function* () {
+          const response = yield* handler(input)
+          const body = yield* Stream.toReadableStreamEffect(response.stream)
+          return HttpClientResponse.fromWeb(
+            input,
+            new Response(body, { status: response.status, headers: response.headers }),
+          )
+        })
+      const error = yield* RequestExecutor.stream(executor, sseRequest, rewrap).pipe(
+        Stream.runForEach((chunk) => Effect.sync(() => chunks.push(chunk))),
+        Effect.flip,
+      )
+
+      expectAIError(error)
+      expect(new TextDecoder().decode(chunks[0])).toBe('data: {"type":"ping"}\n\n')
+      expect(error.message).toBe("Connection lost while reading the response: ECONNRESET: other side closed")
+      expect(error.reason.cause).toBeInstanceOf(TypeError)
+      expect(error.reason.http).toMatchObject({ status: 200 })
+      expect(error.reason).toMatchObject({ _tag: "Transport", operation: "read", code: "ECONNRESET" })
+      expect(isRetryable(error)).toBeTrue()
+    }).pipe(
+      Effect.provide(
+        truncatedStream(
+          ['data: {"type":"ping"}\n\n'],
+          new TypeError("terminated", { cause: systemError("ECONNRESET", "other side closed") }),
+        ),
+      ),
+    ),
+  )
+
+  it.effect("does not report a body read failure without a native cause as a decode error", () =>
+    Effect.gen(function* () {
+      const executor = yield* RequestExecutor.Service
+      const error = yield* RequestExecutor.stream(executor, sseRequest).pipe(Stream.runDrain, Effect.flip)
+
+      expectAIError(error)
+      expect(error.message).toBe("Connection lost while reading the response")
+      expect(error.reason).toMatchObject({ _tag: "Transport", operation: "read", code: undefined })
+      expect(isRetryable(error)).toBeTrue()
+    }).pipe(Effect.provide(truncatedStream(['data: {"type":"ping"}\n\n'], new DOMException("aborted", "AbortError")))),
+  )
+
  it.effect("preserves middleware error messages", () =>
    Effect.gen(function* () {
      const executor = yield* RequestExecutor.Service
@@ -273,10 +331,56 @@ describe("RequestExecutor", () => {
      expectAIError(error)
      expect(error.reason).toMatchObject({ _tag: "InvalidRequest" })
      expect("classification" in error.reason ? error.reason.classification : undefined).toBeUndefined()
-      expect(error.message).toBe("Provider request failed with HTTP 400")
+      expect(error.message).toBe("Provider request failed with HTTP 400: invalid parameter")
    }).pipe(Effect.provide(fixedResponse("invalid parameter", { status: 400 }))),
  )

+  it.effect("shows unrecognized provider error bodies", () =>
+    Effect.gen(function* () {
+      const executor = yield* RequestExecutor.Service
+      const error = yield* executor.execute(request).pipe(Effect.flip)
+
+      expect(error.message).toBe(
+        'Provider request failed with HTTP 422: {"object":"error","message":{"detail":[{"msg":"Input should be less than or equal to 1.5"}]}}',
+      )
+    }).pipe(
+      Effect.provide(
+        fixedResponse('{"object":"error","message":{"detail":[{"msg":"Input should be less than or equal to 1.5"}]}}', {
+          status: 422,
+        }),
+      ),
+    ),
+  )
+
+  it.effect("shows messages from common provider error layouts", () =>
+    Effect.gen(function* () {
+      const executor = yield* RequestExecutor.Service
+      const error = yield* executor.execute(request).pipe(Effect.flip)
+
+      expect(error.message).toBe("Invalid API Key")
+    }).pipe(Effect.provide(fixedResponse('{"detail":"Invalid API Key"}', { status: 401 }))),
+  )
+
+  it.effect("truncates long unrecognized provider error bodies", () =>
+    Effect.gen(function* () {
+      const executor = yield* RequestExecutor.Service
+      const error = yield* executor.execute(request).pipe(Effect.flip)
+
+      expect(error.message).toBe(`Provider request failed with HTTP 400: ${"x".repeat(2000)}…`)
+      expect(error.reason.body).toHaveLength(5000)
+    }).pipe(Effect.provide(fixedResponse("x".repeat(5000), { status: 400 }))),
+  )
+
+  it.effect("does not show HTML error pages", () =>
+    Effect.gen(function* () {
+      const executor = yield* RequestExecutor.Service
+      const error = yield* executor.execute(request).pipe(Effect.flip)
+
+      expect(error.message).toBe("Provider request failed with HTTP 502")
+      expect(error.reason.body).toContain("Bad Gateway")
+    }).pipe(Effect.provide(fixedResponse("<!DOCTYPE html><html><body>Bad Gateway</body></html>", { status: 502 }))),
+  )
+
  it.effect("preserves structured provider messages from large error bodies", () =>
    Effect.gen(function* () {
      const executor = yield* RequestExecutor.Service
@@ -299,27 +403,7 @@ describe("RequestExecutor", () => {
    ),
  )

-  it.effect("reads provider messages sent as a plain error string", () =>
-    Effect.gen(function* () {
-      const executor = yield* RequestExecutor.Service
-      const error = yield* executor.execute(request).pipe(Effect.flip)
-
-      expectAIError(error)
-      expect(error.message).toBe("Your team has no credits for this endpoint")
-    }).pipe(
-      Effect.provide(
-        fixedResponse(
-          JSON.stringify({
-            code: "The caller does not have permission to execute the specified operation",
-            error: "Your team has no credits for this endpoint",
-          }),
-          { status: 403 },
-        ),
-      ),
-    ),
-  )
-
-  it.effect("falls back when structured provider messages are empty", () =>
+  it.effect("shows the body when structured provider messages are empty", () =>
    Effect.gen(function* () {
      const executor = yield* RequestExecutor.Service
      const error = yield* executor.execute(request).pipe(Effect.flip)
@@ -328,7 +412,7 @@ describe("RequestExecutor", () => {
      expect(error.reason).toMatchObject({
        _tag: "InvalidRequest",
      })
-      expect(error.message).toBe("Provider request failed with HTTP 400")
+      expect(error.message).toBe('Provider request failed with HTTP 400: {"error":{"message":"  "}}')
    }).pipe(Effect.provide(fixedResponse('{"error":{"message":"  "}}', { status: 400 }))),
  )

@@ -23,6 +23,7 @@ import { Provider as ProviderSubpath } from "@opencode/ai/provider"
 import {
  AssemblyAI,
  Baseten,
+  BlackForestLabs,
  Cartesia,
  CloudflareAIGateway,
  CloudflareWorkersAI,
@@ -32,14 +33,18 @@ import {
  Fal,
  Fireworks,
  Google,
+  Meta,
  OpenCodeZen,
  OpenAI,
  OpenAICompatible,
  OpenRouter,
+  Replicate,
  Runway,
+  Stability,
  TypeSafeAI,
  VercelAIGateway,
  XAI,
+  ZAI,
 } from "@opencode/ai/providers"
 import {
  OpenAIChat,
@@ -149,14 +154,36 @@ describe("public exports", () => {
    expect(XAI.model).toBeFunction()
    expect(XAI.provider.responses).toBe(XAI.responses)
    expect(XAI.provider.chat).toBe(XAI.chat)
-    expect(XAI.configure({ apiKey: "fixture" }).responses("grok-4.3").route.id).toBe("openai-responses")
-    expect(XAI.configure({ apiKey: "fixture" }).chat("grok-4.3").route.id).toBe("openai-compatible-chat")
+    expect(XAI.configure({ apiKey: "fixture" }).responses("grok-4.3").route.id).toBe("xai-responses")
+    expect(XAI.configure({ apiKey: "fixture" }).chat("grok-4.3").route.id).toBe("xai-chat")
+    expect(OpenAI.configure({ apiKey: "fixture" }).image("gpt-image-2").route.id).toBe("openai-images")
+    expect(OpenAI.provider.image).toBe(OpenAI.image)
+    expect(Google.configure({ apiKey: "fixture" }).image("imagen-4.0-generate-001").route.id).toBe("google-images")
+    expect(Google.provider.image).toBe(Google.image)
+    expect(XAI.configure({ apiKey: "fixture" }).image("grok-imagine-image").route.id).toBe("xai-images")
+    expect(XAI.provider.image).toBe(XAI.image)
+    expect(Fal.configure({ apiKey: "fixture" }).image("fal-ai/flux/dev").route.id).toBe("fal-images")
+    expect(Fal.provider.image).toBe(Fal.image)
+    expect(BlackForestLabs.configure({ apiKey: "fixture" }).image("flux-2-pro").route.id).toBe("bfl-images")
+    expect(BlackForestLabs.provider.image).toBe(BlackForestLabs.image)
+    expect(Replicate.configure({ apiKey: "fixture" }).image("black-forest-labs/flux-schnell").route.id).toBe(
+      "replicate-images",
+    )
+    expect(Replicate.provider.image).toBe(Replicate.image)
+    expect(Stability.configure({ apiKey: "fixture" }).image("sd3.5-large").route.id).toBe("stability-images")
+    expect(Stability.provider.image).toBe(Stability.image)
+    expect(Stability.configure({ apiKey: "fixture" }).upscale().route.id).toBe("stability-upscale")
+    expect(Stability.provider.upscale).toBe(Stability.upscale)
+    expect(Meta.configure({ apiKey: "fixture" }).image("muse-image").route.id).toBe("meta-images")
+    expect(Meta.provider.image).toBe(Meta.image)
+    expect(ZAI.configure({ apiKey: "fixture" }).image("glm-image").route.id).toBe("zai-images")
+    expect(ZAI.provider.image).toBe(ZAI.image)
    expect(XAI.configure({ apiKey: "fixture" }).video("grok-imagine-video-1.5").route.id).toBe("xai-video")
-    expect(XAI.configure({ apiKey: "fixture" }).speech("grok-tts").route.id).toBe("xai-speech")
-    expect(XAI.configure({ apiKey: "fixture" }).transcription("grok-voice-transcribe-2.0").route.kind).toBe("inline")
-    expect(XAI.provider.speech).toBe(XAI.speech)
-    expect(XAI.provider.transcription).toBe(XAI.transcription)
+    expect(XAI.provider.video).toBe(XAI.video)
+    expect(Google.configure({ apiKey: "fixture" }).video("veo-3.1-generate-preview").route.id).toBe("google-video")
+    expect(Google.provider.video).toBe(Google.video)
    expect(Fal.configure({ apiKey: "fixture" }).video("fal-ai/veo3.1").route.id).toBe("fal-video")
+    expect(Fal.provider.video).toBe(Fal.video)
    expect(Runway.configure({ apiKey: "fixture" }).video("gen4.5").route.id).toBe("runway-video")
    expect(Runway.provider.video).toBe(Runway.video)
    expect(OpenAI.configure({ apiKey: "fixture" }).speech("gpt-4o-mini-tts").route.id).toBe("openai-speech")
@@ -170,6 +197,11 @@ describe("public exports", () => {
    expect(Google.configure({ apiKey: "fixture" }).transcription("gemini-3.5-transcribe").route.kind).toBe("stream")
    expect(Deepgram.configure({ apiKey: "fixture" }).transcription("nova-3").route.kind).toBe("inline")
    expect(AssemblyAI.configure({ apiKey: "fixture" }).transcription("universal-3-5-pro").route.kind).toBe("queued")
+    expect(ElevenLabs.configure({ apiKey: "fixture" }).transcription("scribe_v2").route.id).toBe(
+      "elevenlabs-transcription",
+    )
+    expect(ElevenLabs.configure({ apiKey: "fixture" }).transcription("scribe_v2").route.kind).toBe("inline")
+    expect(ElevenLabs.provider.transcription).toBe(ElevenLabs.transcription)
  })

  test("protocol barrels expose supported low-level routes", () => {
@@ -10,7 +10,7 @@
      "usage"
    ],
    "name": "alibaba-chat/qwen-3-7-plus-streams-thinking-disabled",
-    "recordedAt": "2026-09-08T03:10:42.782Z"
+    "recordedAt": "2026-10-02T01:23:37.836Z"
  },
  "interactions": [
    {
@@ -21,14 +21,14 @@
        "headers": {
          "content-type": "application/json"
        },
-        "body": "{\"model\":\"qwen3.7-plus\",\"messages\":[{\"role\":\"user\",\"content\":\"What is 173 multiplied by 219? Reply with only the final integer.\"}],\"stream\":true,\"stream_options\":{\"include_usage\":true},\"max_completion_tokens\":4096,\"enable_thinking\":false}"
+        "body": "{\"model\":\"qwen3.7-plus\",\"messages\":[{\"role\":\"user\",\"content\":[{\"type\":\"text\",\"text\":\"What is 173 multiplied by 219? Reply with only the final integer.\",\"cache_control\":{\"type\":\"ephemeral\"}}]}],\"stream\":true,\"stream_options\":{\"include_usage\":true},\"max_completion_tokens\":4096,\"enable_thinking\":false}"
      },
      "response": {
        "status": 200,
        "headers": {
          "content-type": "text/event-stream;charset=utf-8"
        },
-        "body": "data: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-047bcb67-b193-9a4f-9d77-ea0325be6d7c\",\"created\":1788837041,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\"},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-047bcb67-b193-9a4f-9d77-ea0325be6d7c\",\"choices\":[{\"delta\":{\"content\":\"3\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837041,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-047bcb67-b193-9a4f-9d77-ea0325be6d7c\",\"choices\":[{\"delta\":{\"content\":\"7887\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837041,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-047bcb67-b193-9a4f-9d77-ea0325be6d7c\",\"choices\":[{\"delta\":{\"content\":\"\"},\"index\":0,\"finish_reason\":\"stop\",\"logprobs\":null}],\"created\":1788837041,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"choices\":[],\"created\":1788837041,\"id\":\"chatcmpl-047bcb67-b193-9a4f-9d77-ea0325be6d7c\",\"model\":\"qwen3.7-plus\",\"object\":\"chat.completion.chunk\",\"usage\":{\"completion_tokens\":5,\"prompt_tokens\":32,\"prompt_tokens_details\":{\"cached_tokens\":0,\"text_tokens\":32},\"total_tokens\":37}}\n\ndata: [DONE]\n\n"
+        "body": "data: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-e96a3356-bec9-98ad-bdbe-af00a15ee965\",\"created\":1790904216,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\"},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-e96a3356-bec9-98ad-bdbe-af00a15ee965\",\"choices\":[{\"delta\":{\"content\":\"3\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904216,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-e96a3356-bec9-98ad-bdbe-af00a15ee965\",\"choices\":[{\"delta\":{\"content\":\"7887\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904216,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-e96a3356-bec9-98ad-bdbe-af00a15ee965\",\"choices\":[{\"delta\":{\"content\":\"\"},\"index\":0,\"finish_reason\":\"stop\",\"logprobs\":null}],\"created\":1790904216,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"choices\":[],\"created\":1790904216,\"id\":\"chatcmpl-e96a3356-bec9-98ad-bdbe-af00a15ee965\",\"model\":\"qwen3.7-plus\",\"object\":\"chat.completion.chunk\",\"usage\":{\"completion_tokens\":5,\"prompt_tokens\":32,\"prompt_tokens_details\":{\"cache_creation\":{\"ephemeral_5m_input_tokens\":0},\"cache_creation_input_tokens\":0,\"cache_type\":\"ephemeral\",\"cache_write_tokens\":0,\"cached_tokens\":0,\"text_tokens\":32},\"total_tokens\":37}}\n\ndata: [DONE]\n\n"
      }
    }
  ]
@@ -10,7 +10,7 @@
      "usage"
    ],
    "name": "alibaba-chat/qwen-3-7-plus-streams-thinking-enabled",
-    "recordedAt": "2026-09-08T03:10:54.099Z"
+    "recordedAt": "2026-10-02T01:23:43.998Z"
  },
  "interactions": [
    {
@@ -21,14 +21,14 @@
        "headers": {
          "content-type": "application/json"
        },
-        "body": "{\"model\":\"qwen3.7-plus\",\"messages\":[{\"role\":\"user\",\"content\":\"What is 173 multiplied by 219? Reply with only the final integer.\"}],\"stream\":true,\"stream_options\":{\"include_usage\":true},\"max_completion_tokens\":4096,\"enable_thinking\":true,\"thinking_budget\":1024}"
+        "body": "{\"model\":\"qwen3.7-plus\",\"messages\":[{\"role\":\"user\",\"content\":[{\"type\":\"text\",\"text\":\"What is 173 multiplied by 219? Reply with only the final integer.\",\"cache_control\":{\"type\":\"ephemeral\"}}]}],\"stream\":true,\"stream_options\":{\"include_usage\":true},\"max_completion_tokens\":4096,\"enable_thinking\":true,\"thinking_budget\":1024}"
      },
      "response": {
        "status": 200,
        "headers": {
          "content-type": "text/event-stream;charset=utf-8"
        },
-        "body": "data: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\",\"reasoning_content\":\"\"},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"Thinking\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" Process:\\n\\n1\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\".  **Ident\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"ify the core question\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\":** The user wants\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" to know the product\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" of \"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"173 and\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" 219\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\".\\n2.\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"  **Identify\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" the constraint:** The\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" response must contain *\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"only\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-ec928416-c99b-9f46-9805-cbb37bd7c499\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"* the final integer\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837042,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"cLine truncated
+        "body": "data: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\",\"reasoning_content\":\"\"},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"Thinking\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" Process:\\n\\n1\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\".  Identify\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" the core\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" question: Calculate \"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"173 *\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" 219\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\".\\n2.\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"  Constraint: Reply\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" with *\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"only* the final\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" integer.\\n3\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\".  Perform the\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\" multiplication:\\n\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fe-9ee8-a442-7a3881697fdf\",\"choices\":[{\"delta\":{\"content\":\"\",\"reasoning_content\":\"    *   \"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904218,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.7-plus\",\"id\":\"chatcmpl-70ed9ebc-62fLine truncated
      }
    }
  ]
@@ -1,9 +1,15 @@
 {
  "version": 1,
  "metadata": {
-    "tags": ["prefix:alibaba-chat", "provider:alibaba", "protocol:alibaba-chat", "region:ap-southeast-1", "image"],
+    "tags": [
+      "prefix:alibaba-chat",
+      "provider:alibaba",
+      "protocol:alibaba-chat",
+      "region:ap-southeast-1",
+      "image"
+    ],
    "name": "alibaba-chat/qwen-3-8-flash-reads-image-bytes",
-    "recordedAt": "2026-09-08T03:10:55.585Z"
+    "recordedAt": "2026-10-02T01:23:45.179Z"
  },
  "interactions": [
    {
@@ -14,14 +20,14 @@
        "headers": {
          "content-type": "application/json"
        },
-        "body": "{\"model\":\"qwen3.8-flash\",\"messages\":[{\"role\":\"user\",\"content\":[{\"type\":\"text\",\"text\":\"Read the three words in this image. Reply with only the words in order.\"},{\"type\":\"image_url\",\"image_url\":{\"url\":\"data:image/png;base64,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 truncated
+        "body": "{\"model\":\"qwen3.8-flash\",\"messages\":[{\"role\":\"user\",\"content\":[{\"type\":\"text\",\"text\":\"Read the three words in this image. Reply with only the words in order.\",\"cache_control\":{\"type\":\"ephemeral\"}},{\"type\":\"image_url\",\"image_url\":{\"url\":\"data:image/png;base64,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 truncated
      },
      "response": {
        "status": 200,
        "headers": {
          "content-type": "text/event-stream;charset=utf-8"
        },
-        "body": "data: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-2952c591-6e41-90ec-81ca-894b8e84f0aa\",\"created\":1788837054,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\"},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-2952c591-6e41-90ec-81ca-894b8e84f0aa\",\"choices\":[{\"delta\":{\"content\":\"j\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837054,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-2952c591-6e41-90ec-81ca-894b8e84f0aa\",\"choices\":[{\"delta\":{\"content\":\"iggling restroom prison\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837054,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-2952c591-6e41-90ec-81ca-894b8e84f0aa\",\"choices\":[{\"delta\":{\"content\":\"\"},\"index\":0,\"finish_reason\":\"stop\",\"logprobs\":null}],\"created\":1788837054,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"choices\":[],\"created\":1788837054,\"id\":\"chatcmpl-2952c591-6e41-90ec-81ca-894b8e84f0aa\",\"model\":\"qwen3.8-flash\",\"object\":\"chat.completion.chunk\",\"usage\":{\"completion_tokens\":6,\"completion_tokens_details\":{},\"prompt_tokens\":110,\"prompt_tokens_details\":{\"cached_tokens\":0,\"image_tokens\":82,\"text_tokens\":28},\"total_tokens\":116}}\n\ndata: [DONE]\n\n"
+        "body": "data: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-69096b14-dda4-93fc-a838-0e7640f71dfa\",\"created\":1790904224,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\"},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-69096b14-dda4-93fc-a838-0e7640f71dfa\",\"choices\":[{\"delta\":{\"content\":\"j\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904224,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-69096b14-dda4-93fc-a838-0e7640f71dfa\",\"choices\":[{\"delta\":{\"content\":\"igg\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904224,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-69096b14-dda4-93fc-a838-0e7640f71dfa\",\"choices\":[{\"delta\":{\"content\":\"ling restroom prison\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904224,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-flash\",\"id\":\"chatcmpl-69096b14-dda4-93fc-a838-0e7640f71dfa\",\"choices\":[{\"delta\":{\"content\":\"\"},\"index\":0,\"finish_reason\":\"stop\",\"logprobs\":null}],\"created\":1790904224,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"choices\":[],\"created\":1790904224,\"id\":\"chatcmpl-69096b14-dda4-93fc-a838-0e7640f71dfa\",\"model\":\"qwen3.8-flash\",\"object\":\"chat.completion.chunk\",\"usage\":{\"completion_tokens\":6,\"completion_tokens_details\":{},\"prompt_tokens\":123,\"prompt_tokens_details\":{\"cached_tokens\":0,\"image_tokens\":82,\"text_tokens\":41},\"total_tokens\":129}}\n\ndata: [DONE]\n\n"
      }
    }
  ]
@@ -10,7 +10,7 @@
      "tool-choice"
    ],
    "name": "alibaba-chat/qwen-3-8-max-obeys-named-tool-choice",
-    "recordedAt": "2026-09-08T03:10:56.576Z"
+    "recordedAt": "2026-10-02T01:23:46.235Z"
  },
  "interactions": [
    {
@@ -21,14 +21,14 @@
        "headers": {
          "content-type": "application/json"
        },
-        "body": "{\"model\":\"qwen3.8-max\",\"messages\":[{\"role\":\"user\",\"content\":\"Find the current weather in Paris.\"}],\"tools\":[{\"type\":\"function\",\"function\":{\"name\":\"get_weather\",\"description\":\"Get weather in a city\",\"parameters\":{\"type\":\"object\",\"properties\":{\"city\":{\"type\":\"string\",\"enum\":[\"Paris\"]}},\"required\":[\"city\"]}}}],\"tool_choice\":{\"type\":\"function\",\"function\":{\"name\":\"get_weather\"}},\"stream\":true,\"stream_options\":{\"include_usage\":true},\"reasoning_effort\":\"none\",\"max_completion_tokens\":4096}"
+        "body": "{\"model\":\"qwen3.8-max\",\"messages\":[{\"role\":\"user\",\"content\":[{\"type\":\"text\",\"text\":\"Find the current weather in Paris.\",\"cache_control\":{\"type\":\"ephemeral\"}}]}],\"tools\":[{\"type\":\"function\",\"function\":{\"name\":\"get_weather\",\"description\":\"Get weather in a city\",\"parameters\":{\"type\":\"object\",\"properties\":{\"city\":{\"type\":\"string\",\"enum\":[\"Paris\"]}},\"required\":[\"city\"]}}}],\"tool_choice\":{\"type\":\"function\",\"function\":{\"name\":\"get_weather\"}},\"stream\":true,\"stream_options\":{\"include_usage\":true},\"reasoning_effort\":\"none\",\"max_completion_tokens\":4096}"
      },
      "response": {
        "status": 200,
        "headers": {
          "content-type": "text/event-stream;charset=utf-8"
        },
-        "body": "data: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\",\"tool_calls\":[{\"index\":0,\"id\":\"call_4eb823cb28d141c8befbb331\",\"type\":\"function\",\"function\":{\"name\":\"get_weather\",\"arguments\":\"\"}}]},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"index\":0,\"id\":\"\",\"type\":\"function\",\"function\":{\"arguments\":\"\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"{\\\"city\\\": \\\"Paris\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"\\\"\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"}\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"choices\":[{\"delta\":{\"tool_calls\":[{\"function\":{\"arguments\":\"\"},\"index\":0,\"id\":null,\"type\":\"function\"}],\"content\":\"\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"choices\":[{\"delta\":{},\"index\":0,\"finish_reason\":\"stop\",\"logprobs\":null}],\"created\":1788837055,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"choices\":[],\"created\":1788837055,\"id\":\"chatcmpl-681dfa5a-ae08-98da-ba0d-4b1d0d364b6b\",\"model\":\"qwen3.8-max\",\"object\":\"chat.completion.chunk\",\"usage\":{\"completion_tokens\":19,\"prompt_tokens\":288,\"prompt_tokens_details\":{\"cached_tokens\":0,\"text_tokens\":288},\"total_tokens\":307}}\n\ndata: [DONE]\n\n"
+        "body": "data: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null,\"choices\":[{\"logprobs\":null,\"index\":0,\"delta\":{\"content\":\"\",\"role\":\"assistant\",\"tool_calls\":[{\"index\":0,\"id\":\"call_3b220e76d7d946f498d026c0\",\"type\":\"function\",\"function\":{\"name\":\"get_weather\",\"arguments\":\"\"}}]},\"finish_reason\":null}]}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"index\":0,\"id\":\"\",\"type\":\"function\",\"function\":{\"arguments\":\"\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"{\\\"city\\\": \\\"Paris\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"\\\"\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"}\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"choices\":[{\"delta\":{\"content\":\"\",\"tool_calls\":[{\"type\":\"function\",\"index\":0,\"function\":{\"arguments\":\"\"}}]},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"choices\":[{\"delta\":{\"tool_calls\":[{\"function\":{\"arguments\":\"\"},\"index\":0,\"id\":null,\"type\":\"function\"}],\"content\":\"\"},\"index\":0,\"finish_reason\":null,\"logprobs\":null}],\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"model\":\"qwen3.8-max\",\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"choices\":[{\"delta\":{},\"index\":0,\"finish_reason\":\"stop\",\"logprobs\":null}],\"created\":1790904225,\"object\":\"chat.completion.chunk\",\"usage\":null}\n\ndata: {\"choices\":[],\"created\":1790904225,\"id\":\"chatcmpl-bf693c1a-0554-935b-ab28-cda0a18a27ff\",\"model\":\"qwen3.8-max\",\"object\":\"chat.completion.chunk\",\"usage\":{\"completion_tokens\":19,\"prompt_tokens\":288,\"prompt_tokens_details\":{\"cache_creation\":{\"ephemeral_5m_input_tokens\":0},\"cache_creation_input_tokens\":0,\"cache_type\":\"ephemeral\",\"cache_write_tokens\":0,\"cached_tokens\":0,\"text_tokens\":288},\"total_tokens\":307}}\n\ndata: [DONE]\n\n"
      }
    }
  ]