arcane-os 0.69.0 → 0.71.0

package/docs/reference/runtime-modules.md

@@ -477,6 +477,13 @@ the caller's ready DBOPFS instance; the separate JSON-compatible `identity`
 adds application-owned semantic context. Reuse compares complete inputs rather
 than an SDK version alone.
 
+Only missing audio joins TTS activation. Each preparation supplies its own
+cancellation lifetime to the existing provider runtime. The last pending
+activation interest cancels the owned load; other preparation/playback or
+joining direct-load interests preserve it. Cache-only preparation never loads
+the model, and cancelling after activation does not unload an already-ready
+model. Provider work owns actual cancellation and cleanup completion.
+
 `state` is `queued`, `preparing`, `ready`, `error`, or `cancelled`.
 `onState({state,completed,total,segments,error})` synchronously observes
 preparation progress. `ready` resolves the complete record after every segment
@@ -486,8 +493,8 @@ the retained MIME type. Preparation failure rejects; cancellation rejects as
 `AbortError` and preserves successfully stored segments. Matching pending
 requests share synthesis only on the same AI instance and storage group;
 cancelling one handle does not cancel another active matching caller.
-Preparation cancellation prevents later synthesis, but an already-started
-shared provider load/unmute has no per-preparation signal and may finish.
+Preparation cancellation prevents later synthesis and releases its activation
+interest. The provider's operation remains responsible for actual cleanup.
 
 `playPreparedTTS(prepared,{signal,onState})` can attach immediately. Its returned
 `{state,error,finished,pause(),resume(),stop()}` handle schedules segments in
@@ -1024,9 +1031,20 @@ The singleton exposes read-only `protocol`, `configured`, and `speechMuted`;
 `supportsTranscriptionCapture(providerId=null)`;
 `createTranscriptionCapture(options={})`;
 `synthesize(payload,options={},preparation={})`; and
-`setSpeechMuted(muted)`. Provider payloads must be data-only; callbacks,
+`setSpeechMuted(muted,{signal}={})`. Provider payloads must be data-only; callbacks,
 accessors, symbols, and cycles are rejected at the provider boundary.
 
+`setSpeechMuted(false,{signal})` shares pending activation while retaining an
+independent interest for each caller. Aborting one signal rejects only that
+caller's wait. When no pending interests remain, abort reaches the owned load.
+No-signal callers and direct `load('tts',...)` calls joining that activation
+also retain it. A rejected load preserves its complete error for the remaining
+callers. An already-aborted signal rejects before changing mute intent.
+Explicit `setSpeechMuted(true)` still cancels shared activation and unloads TTS;
+once accepted, that mute cleanup runs to completion rather than being undone
+by cancellation of its caller's signal. Completed activation is not unloaded
+merely because a former observer later aborts.
+
 Selection options admit `localOnly=false`; inspection admits
 `{localOnly=false,signal=null}`; startup admits
 `{startLanguageModel=true,startMuted=true,startTranscription=false,signal=null}` (including an omitted
@@ -2203,6 +2221,59 @@ default `DBOPFS`; installs `window.dbopfs`, emits `dbopfs-ready`; table/file/bac
 
 Exact exports: `DBOPFS_EVENT_TYPES`, `DBOPFS_REASONS`, `default`.
 
+`subscribeChanges(listener, {signal, once})` observes
+`DBOPFS_EVENT_TYPES.change` (`dbopfs-change`) through the existing canonical
+`arcane-os/event-manager` source. It returns an idempotent unsubscribe function
+with `.dispose()`; aborting the supplied signal also removes the subscription.
+Each listener receives an event whose `detail` contains:
+
+| Field | Meaning |
+| --- | --- |
+| `applicationId`, `storagePath` | Existing application identity and `apps/<application-id>` OPFS scope. |
+| `tableName`, `directoryName` | Logical table and physical directory; `memories` / `memory` retain their existing alias. |
+| `fileName` | Complete record filename, or `null` for a removed table. |
+| `action` | `write`, `append`, `delete`, or `table-delete`. A write does not claim whether a record was newly created. |
+| `originId`, `sequence`, `changeId` | Per-document origin and incrementing committed occurrence, shared by local and remote notifications for that occurrence. These are correlation fields, separate from realm-local canonical event-source IDs. |
+| `remote` | `false` at the writing DBOPFS source, `true` when received from another live document. |
+
+A write is published after writable-stream `close()` succeeds, or after the
+existing OPFS worker acknowledges its completed write/flush/close. A removal
+is published after OPFS removes the record or table. The affected cached record
+or table, including alias keys and removed-table handles, is invalidated before
+listeners run. A read already in progress may still resolve its earlier
+snapshot; it cannot repopulate the cache after an intervening delivered change.
+Notification metadata does not contain, rewrite, or reread record bodies.
+
+`set()` and raw `writeFile()` publish one occurrence per completed write;
+`setMany()`, `deleteMany()`, clear operations, and restore retain their existing
+per-record results and publish only their actual successful mutations. A
+partial batch is not a transaction: successful records are announced even when
+siblings fail. Missing records, absent/non-empty empty-table removal, failed
+writes/removals, and reads emit no committed-change event. `deleteTable()` keeps
+its existing return/error-reporting behavior; its returned `true` alone does
+not imply that a committed-change event was emitted.
+
+Within one document, subscribe through this source or the canonical event
+manager, rather than consuming both as separate changes. Across documents,
+DBOPFS owns one native `BroadcastChannel` named from the application storage
+path. Native storage-partition matching plus that path selects the receiving
+documents. A remote receiver invalidates its affected cache before publishing
+locally and never rebroadcasts the occurrence. There is no echo back to the
+writing DBOPFS source. Ordinary storage success remains success if a transport
+notification fails; the notification error is reported separately.
+
+Subscriptions are live-only: subscribe before the application's initial read,
+then let the application map the table/file/action to its own records and UI.
+DBOPFS performs no polling, initial full-table read, automatic picker refresh,
+or event-history replay. The channel lasts for the owning page, independent of
+individual listener lifetimes, so a document that only writes still announces
+its commits. `pagehide` closes the channel; `pageshow` after suspension reopens
+it and clears only that document's in-memory caches. Resuming a page does not
+fabricate a storage mutation. Applications own any resume-time view refresh.
+Missed events during suspension, process termination, or realm loss are not
+replayed; the stored records remain the authority. Direct OPFS writes made
+outside DBOPFS are not observed by this contract.
+
 `createCompressedPNG({tableNames, additionalTables, signal})` returns an
 `image/png` Blob without downloading or writing records. Omitted `tableNames`
 selects all discovered saved tables; `[]` selects none. Named absent tables
@@ -2244,16 +2315,35 @@ unreadable row remains in its original string form so the owning application
 can display, diagnose, or recover it without silent data loss. DOM and storage
 errors remain observable. The logical `memories` table continues to map to the
 physical `memory` directory, and successful or already-absent empty-table
-removal invalidates both alias and cached-handle state. Transport: OPFS,
-DBOPFSWorker, Compression Streams.
+removal invalidates both alias and cached-handle state. Cross-document change
+delivery requires native `BroadcastChannel` in the same origin and storage
+partition, with the same application scope; separate WebView profiles do not
+share this transport. If it is unavailable, local change subscriptions and
+ordinary DBOPFS storage continue, without cross-document notification or a
+polling replacement. Transport: OPFS, DBOPFSWorker, Compression Streams, native
+BroadcastChannel, and the existing event-pubsub-backed canonical event owner.
 [Deep protocol details](protocols.md).
 
 ### Example
 
 ```javascript
-import * as module from '/arcane/modules/DBOPFS.js';
+import '/arcane/modules/DBOPFS.js';
+
+await dbopfs.readyPromise;
+const lifetime = new AbortController();
+const stop = dbopfs.subscribeChanges(function showCommittedParade(event) {
+  const change = event.detail;
+  if (change.tableName === 'moon_bus') {
+    console.log(change.action, change.fileName, change.remote);
+  }
+}, {signal: lifetime.signal});
 
-console.log(Object.keys(module));
+await dbopfs.set('moon_bus', 'squid-parade.json', {
+  destination: 'The moon', passengers: ['Captain Squid', 'One confused lobster']
+});
+// Another live page using this app/storage scope receives the same occurrence.
+// The application can read that one record when its view needs the body.
+stop(); // lifetime.abort() is the equivalent subscription cleanup.
 ```
 
 ## DBOPFSDocumentLibrary.js

package/docs/reference/sdk-api.md

@@ -136,7 +136,7 @@ Root applications import `AI` through `arcane-os/ai` in the managed browser map.
 `ai.prepareTTSPlayback(payload,signal)` silently prepares real Blob audio or a
 native speech descriptor for playback. Native browser voices expose no audio
 file; use this path or shared streaming/SpeechPlayback for native output.
-`ai.prepareTTS({parts,storage,identity,signal,onState})` prepares complete
+`ai.prepareTTS({parts,storage,identity,textFormat,signal,onState})` prepares complete
 punctuation segments without playing them and optionally saves/reuses their
 audio through application-owned DBOPFS. Its immediate handle exposes ordered
 segments, preparation state, `ready`, `getAudio(index)`, and `cancel()`.
@@ -149,6 +149,17 @@ for complete inputs, storage ownership, progress, errors, and cancellation,
 and [AI.js](runtime-modules.md#aijs) for the managed runtime contract.
 These are AI instance methods, not new npm export names.
 
+Complete synthesis calls can select `textFormat:'plain'` to preserve supplied
+speech text, including repeated formatting marks and surrounding whitespace.
+This applies to `ai.fetchTTS(payload,signal)`, audio-file preparation through
+`ai.prepareTTSPlayback(payload,signal)`, durable `ai.prepareTTS(...)`, provider
+runtime TTS requests and direct Kokoro synthesis. The SDK consumes the control
+before engine dispatch; durable preparation retains the existing punctuation
+segmentation. Omitted or other values retain automatic speech cleanup.
+Streaming `streamTTS`, `finishTTS` and shared `SpeechPlayback` retain their
+automatic cleanup; this option does not change upstream engine capacity.
+See [exact complete synthesis input](ai/browser-speech.md#exact-complete-synthesis-input).
+
 ## Shared operation contract
 
 High-level operations accept one options object. Common fields are `workspaceRoot`, `appId`, `target`, `signal`, and `onEvent`; only fields meaningful to that operation are consumed. Long work acknowledges first, owns its child tasks, keeps event delivery ordered and backpressured, emits progress or heartbeats, observes cancellation where safe, and rejects with `ArcaneError` on failure.
@@ -7530,6 +7541,12 @@ published SDK/runtime dependencies and the canonical Core entry/client into a
 fresh output directory. Service factories are imported only when a real native
 host later launches the generated Core entry, never during packaging.
 
+An optional `appDescriptor.native.window` record carries `width`, `height` and
+`resizable` into `manifest.window`. Dimensions are positive integral logical
+client units; `resizable` is an optional boolean. Omitted fields remain omitted.
+The portable payload records the selection; the selected platform host owns
+its window behavior. See [initial native window size](core-native-packaging.md#initial-native-window-size).
+
 ### Availability and normalization
 
 **Node; produces a portable payload rather than an executable host.** Existing
@@ -7641,6 +7658,16 @@ the selected output root. There is no compiler or OS checkout requirement for
 application builds. The build machine needs its normal Node/npm toolchain and
 `tar`; execution needs Windows x64, .NET Framework 4.6.2 or later and WebView2.
 
+The matching Windows host consumes `appDescriptor.native.window` through
+`manifest.window`. Optional `width` and `height` select initial client dimensions
+in 96-DPI logical units; the host scales them for its actual DPI and fits the
+outer window within the current screen's usable area. Optional `resizable:false`
+selects a fixed-size border and disables maximizing. Omitted fields preserve
+the existing host defaults. Initial sizing preserves a window already restored
+to a non-normal state and does not add saved-bounds storage. The initial-fit
+logic is not reapplied during later user resizing. See [initial native window size](core-native-packaging.md#initial-native-window-size)
+for a complete descriptor example and the other-platform boundary.
+
 The returned `arcane-native-builder/1` provider exposes `describe`, `doctor`,
 `prepare`, `build`, `verify` and `run`. Build returns `{app,target,manifest}`
 with the actual directory at `target.rootDir`. Verification reads the artifact;
@@ -8844,13 +8871,14 @@ occurrences use source `ai-model-controller`; every load or unload owns one
 non-null operation ID shared by its state and progress occurrences. Public
 progress retains the exact selected `modelId`, `phase`, state, and nested
 file-progress field names. Browser-WASM download progress preserves
-`completed`, `total`, and `unit:'files'` and adds `loadedBytes`, `totalBytes`,
-`remainingBytes`, `bytesPerSecond`, `etaSeconds`, `activeTransfers`,
-`transferLimit`, and `transferMode`. Unknown totals, rates, remaining values,
-and ETAs are `null`; applications format the raw byte and second measures for
-display. Chunk-driven transfer changes are coalesced on a 250 ms cadence, while
-plan, active-worker, and completion boundaries publish immediately and the existing five-second
-heartbeat repeats the latest record.
+`completed`, `total`, and `unit:'files'`, with complete per-file `members`
+reporting stored shards. A member's `total` remains `null` until its complete
+response is stored, unless the HTTP range plan already establishes its shard
+total. `activeTransfers`, `transferLimit`, and `transferMode` describe the
+active download work. Closed shards, active-worker changes, and completion
+boundaries publish semantic progress; the existing heartbeat repeats the latest
+record. These fields describe files and shards, without byte-based display
+requirements or estimated transfer rates.
 Accessor-bearing provider status
 and malformed progress reject with `ARCANE_AI_PROVIDER_STATUS_INVALID` and
 `ARCANE_AI_PROVIDER_PROGRESS_INVALID`. After disposal, lifecycle operations fail with
@@ -9443,8 +9471,8 @@ console.log(provider.status().state); // unloaded
 ### Overview
 
 Adapts an existing DBOPFS instance into a complete multi-model, ordered
-multi-file cache without renaming its public methods. Each completed whole
-member or Range part is committed independently. An interrupted set
+multi-file cache without renaming its public methods. Every new full response
+or Range transfer is committed in ordered parts. An interrupted set
 is not a cache hit, but its completed members remain available to the next
 install so only missing work is fetched.
 
@@ -9462,25 +9490,26 @@ workers
 when the probe returns a usable total from exposed `Content-Range` or optional
 descriptor `bytes`. The mutable
 result contains `kind`, `tableName`, `downloadConcurrency`, the original
-`adapter`, and `ready`, `install`, `ensure`, and `remove`. `ensure()` returns
+`adapter`, and `ready`, `install`, `ensure`, `remove`, and `fetchResource`. `ensure()` returns
 `files` and `cache:'cached'|'installed'`. It preserves the complete model content.
 `offline:true` never downloads and rejects a miss with
 `ARCANE_AI_MODEL_OFFLINE_MISS`.
 
-When supplied, `onProgress` receives existing ordered file progress plus
-`loadedBytes`, `totalBytes`, `remainingBytes`, `bytesPerSecond`, `etaSeconds`,
-`activeTransfers`, `transferLimit`, and `transferMode`. `activeTransfers` is the
-current transfer-worker count. Transfer mode is
-`probing`, `files`, `ranges`, or `single`. Unknown aggregate totals, remaining
-values, rates, and ETAs are `null`. Chunk-driven changes are coalesced on a
-250 ms cadence, with immediate records at download start, plan or total
-discovery, active-worker changes, and completion. `loadedBytes` includes current staged writes and
-restored completed members or Range parts; bytes from a failed partial write
-are removed from the total when that entry is discarded. A complete set of
-declared member `bytes` values
-makes split-model total and remaining progress available from the start;
-otherwise aggregate-dependent fields remain `null` until they are honestly
-known.
+When supplied, `onProgress` receives ordered file progress and `members`
+records with each file's `name`, committed `completed` parts, `total`, and
+`unit:'shards'`. Streaming totals remain `null` until the response finishes;
+previously complete files identify `state:'cached'`. `activeTransfers` is the
+current transfer-worker count, `transferLimit` is selected concurrency, and
+`transferMode` is `probing`, `files`, `ranges`, or `single`. Closed-part and
+lifecycle changes publish promptly, with an existing 250 ms heartbeat. No
+byte progress, transfer rate, or estimated completion time is produced.
+
+`fetchResource(input,{signal,onProgress,...requestOptions})` returns
+`{file,status,statusText,headers,url,redirected}` for a complete stored Fetch
+response. SDK browser model loaders use this shared owner for configuration,
+tokenizer, graph, external-data, and other fetched resources. Successful
+complete GET resources may be reused; supported HTTP ranges resume closed
+parts after interruption. Request framing stays at the transport boundary.
 
 ### Availability and normalization
 
@@ -9491,9 +9520,9 @@ resumable Range parts. For any source member, a followed redirect that
 turns the first `bytes=0-0` probe into `200` triggers one direct probe of the
 final URL. Confirmed support cancels the original body and starts the Range
 pool; refusal keeps the original response as the ordinary single-fetch path,
-with readable `Content-Length` used for progress. A missing `Content-Range` can
+without requiring a readable `Content-Length`. A missing `Content-Range` can
 use optional descriptor `bytes` for Range planning; without either total, the
-store falls back to one full fetch. A confirmed source member is divided into
+store uses one full fetch streamed into ordered persistent parts. A confirmed source member is divided into
 deterministic OPFS Range parts of roughly 4 MB each, up to 4,096 parts. A
 monolithic source may use at most
 `downloadConcurrency` Range workers; split sources keep one Range worker per
@@ -9502,7 +9531,7 @@ in order as one logical Blob. A later non-206, contradictory exposed Range
 response, or incorrectly framed body fails rather than assembling partial
 content. On failure or cancellation, peers settle while completed shards and
 Range parts remain available for retry; an unfinished active part restarts after
-refresh. A persisted Range part whose length does not match its requested HTTP
+refresh. Existing whole-file caches remain readable without migration. A persisted Range part whose length does not match its requested HTTP
 frame is discarded and fetched again. A complete whole member supersedes its
 redundant current Range fragments. Cleanup failure is warned without
 hiding the usable model. Wllama requires every selected model Blob or File to
@@ -9784,7 +9813,7 @@ retains the complete list of deletions already completed.
 
 Adapts an existing DBOPFS instance into a complete caller-selected speech
 runtime/model store. It serializes each selected model with an exclusive Web
-Lock, removes partial state after failure, stores ordinary mutable selection
+Lock, preserves complete peer files and closed parts after failure, stores ordinary mutable selection
 metadata before content, and treats a changed file inventory or source mapping
 as a cache miss that is downloaded again.
 
@@ -9794,7 +9823,7 @@ as a cache miss that is downloaded again.
 createDbopfsSpeechArtifactStore({ dbopfs, tableName='arcane_ai_browser_speech', fetchImpl=null, objectUrlFactory=null }={})
 ```
 
-The mutable result is `{protocol,tableName,prepare,remove}`.
+The mutable result is `{protocol,tableName,prepare,remove,fetchResource}`.
 `prepare(authority,{signal,onProgress,offline=false,security})` uses a complete
 compatible cache or downloads every declared file, preserves complete content
 and materializes object URLs, then returns
@@ -9804,9 +9833,18 @@ needs those URLs. `offline:true` never fetches and rejects a miss with
 authority's files and selection metadata. Optional `onProgress` receives
 preparation and download records with `phase`, `stage`, `message`, and the
 current `file` when available. File operations report actual `completed` and
-`total` inventory counts with `unit:'files'`; module preparation uses
+`total` inventory counts with `unit:'files'`, and `resource` carries the current
+closed-part progress with `unit:'shards'`; module preparation uses
 `total:null`. Cancellation stops progress along with the owning operation.
 
+New files stream into ordered DBOPFS parts with no required response length.
+Existing complete whole files stay readable without migration.
+`fetchResource(input,{signal,onProgress,...requestOptions})` returns
+`{file,status,statusText,headers,url,redirected}` and gives the dedicated
+speech Worker the same complete resource storage owner for runtime fetches.
+Native ESM imports and unmapped classic Worker internals retain their platform
+loader; they are not claimed as intercepted HTTP fetches.
+
 ### Availability and normalization
 
 **Browser with a ready DBOPFS instance, OPFS, Web Locks, Fetch or an injected

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "arcane-os",
-  "version": "0.69.0",
+  "version": "0.71.0",
   "description": "Arcane OS JavaScript SDK, project-local CLI, browser runtime, and repository-portable application packager.",
   "type": "module",
   "main": "./src/index.mjs",
@@ -123,7 +123,7 @@
     "test": "npm run test:unit && npm run test:functional && npm run test:integration && npm run test:regression",
     "test:release": "node ./bin/arcane-test.mjs test/npm-release.test.mjs",
     "test:unit": "node ./bin/arcane-test.mjs test/app-descriptor.test.mjs test/app-schema.test.mjs test/app-selection.test.mjs test/contracts.test.mjs test/conversation-closing-report.test.mjs test/doctor.test.mjs test/mail-credentials.test.mjs test/mail-outbox.test.mjs test/mail-public-api.test.mjs test/mail-send.test.mjs test/mail-transport.test.mjs test/native-windows-icon.test.mjs test/targets.test.mjs test/workspace-operation-lock.test.mjs",
-    "test:functional": "node ./bin/arcane-test.mjs test/ai-ollama-readiness.test.mjs test/core-ollama-lifecycle.test.mjs test/ai-request-model.test.mjs test/ai-request-retry.test.mjs test/ai-stream-cancellation.test.mjs test/browser-decisions.test.mjs test/browser-speech-providers.test.mjs test/browser-wasm-gpu-notice.test.mjs test/browser-wasm-download-resume.test.mjs test/chat-input.test.mjs test/chat-stream-speech.test.mjs test/chat-transcript-scroll.test.mjs test/chat-transient-retention.test.mjs test/cli.test.mjs test/continuous-voice-capture.test.mjs test/codex-app-server.test.mjs test/mcp-stdio-server.test.mjs test/core-classic-source.test.mjs test/core-client.test.mjs test/core-client-installation.test.mjs test/core-ai-installation.test.mjs test/core-llama-lifecycle.test.mjs test/core-development.test.mjs test/core-host.test.mjs test/core-model-assets.test.mjs test/core-native-services.test.mjs test/core-packaged-web.test.mjs test/core-runtime.test.mjs test/core-stdio.test.mjs test/dbopfs-document-library.test.mjs test/dbopfs-png.test.mjs test/dbopfs.test.mjs test/dev-server.test.mjs test/dev-pwa.test.mjs test/dom-event-instrumentation.test.mjs test/event-manager.test.mjs test/events.test.mjs test/file-preview.test.mjs test/import-map.test.mjs test/local-image-editing.test.mjs test/mail-cli.test.mjs test/mail-runtime.test.mjs test/mail-server.test.mjs test/markdown-editor.test.mjs test/markdown-media.test.mjs test/markdown-source-map.test.mjs test/modal.test.mjs test/native-decisions.test.mjs test/native-onnx.test.mjs test/native-module-packaging.test.mjs test/native-module-resource-references.test.mjs test/packaging.test.mjs test/pwa-packaging.test.mjs test/pwa-client.test.mjs test/pwa-install.test.mjs test/pwa-worker.test.mjs test/persistent-ai-chat-session.test.mjs test/reference-completeness.test.mjs test/rendered-print.test.mjs test/runtime-api-behavior.test.mjs test/runtime.test.mjs test/scaffold.test.mjs test/speech-playback.test.mjs test/site.test.mjs test/theme-manager.test.mjs test/twin-cloud-images.test.mjs test/twin-cloud.test.mjs test/websocket-client-import-map.test.mjs test/websocket.test.mjs test/update-check.test.mjs test/user-explicit-updates.test.mjs",
+    "test:functional": "node ./bin/arcane-test.mjs test/ai-ollama-readiness.test.mjs test/core-ollama-lifecycle.test.mjs test/ai-request-model.test.mjs test/ai-request-retry.test.mjs test/ai-stream-cancellation.test.mjs test/browser-decisions.test.mjs test/browser-speech-providers.test.mjs test/browser-speech-artifacts.test.mjs test/browser-speech-resource-bridge.test.mjs test/browser-wasm-gpu-notice.test.mjs test/browser-wasm-download-resume.test.mjs test/chat-input.test.mjs test/chat-stream-speech.test.mjs test/chat-transcript-scroll.test.mjs test/chat-transient-retention.test.mjs test/cli.test.mjs test/continuous-voice-capture.test.mjs test/codex-app-server.test.mjs test/mcp-stdio-server.test.mjs test/core-classic-source.test.mjs test/core-client.test.mjs test/core-client-installation.test.mjs test/core-ai-installation.test.mjs test/core-llama-lifecycle.test.mjs test/core-development.test.mjs test/core-host.test.mjs test/core-shared-host.test.mjs test/core-model-assets.test.mjs test/core-native-services.test.mjs test/core-packaged-web.test.mjs test/core-runtime.test.mjs test/core-stdio.test.mjs test/dbopfs-committed-events.test.mjs test/dbopfs-document-library.test.mjs test/dbopfs-model-resources.test.mjs test/dbopfs-png.test.mjs test/dbopfs.test.mjs test/dev-server.test.mjs test/dev-pwa.test.mjs test/dom-event-instrumentation.test.mjs test/event-manager.test.mjs test/events.test.mjs test/file-preview.test.mjs test/import-map.test.mjs test/local-image-editing.test.mjs test/mail-cli.test.mjs test/mail-runtime.test.mjs test/mail-server.test.mjs test/markdown-editor.test.mjs test/markdown-media.test.mjs test/markdown-source-map.test.mjs test/modal.test.mjs test/model-resource-bridge.test.mjs test/native-decisions.test.mjs test/native-onnx.test.mjs test/native-module-packaging.test.mjs test/native-module-resource-references.test.mjs test/packaging.test.mjs test/pwa-packaging.test.mjs test/pwa-client.test.mjs test/pwa-install.test.mjs test/pwa-worker.test.mjs test/persistent-ai-chat-session.test.mjs test/reference-completeness.test.mjs test/rendered-print.test.mjs test/runtime-api-behavior.test.mjs test/runtime.test.mjs test/scaffold.test.mjs test/speech-playback.test.mjs test/site.test.mjs test/theme-manager.test.mjs test/twin-cloud-images.test.mjs test/twin-cloud.test.mjs test/websocket-client-import-map.test.mjs test/websocket.test.mjs test/update-check.test.mjs test/user-explicit-updates.test.mjs",
     "test:integration": "node ./bin/arcane-test.mjs test/installed-package-runtime.test.mjs test/root-app-layout.test.mjs test/integrated-shared.test.mjs test/integrated-workspace.test.mjs test/mail-browser.test.mjs test/native-macos-provider.test.mjs test/native-plan.test.mjs test/native-portable-provider.test.mjs test/native-provider-loader.test.mjs test/npm-release.test.mjs test/release-bundle.test.mjs test/release-capability-smoke.test.mjs test/shared-payload-batch.test.mjs test/tarball.test.mjs test/browser-wasm-cpu.test.mjs test/wllama-webgpu-runtime.test.mjs",
     "test:regression": "node ./bin/arcane-test.mjs test/channel-workflows.test.mjs test/html-import-registration.test.mjs test/html-import-packaged-script.test.mjs test/logging-regression.test.mjs test/markdown-speech.test.mjs test/prepared-speech.test.mjs test/native-provider-generation.test.mjs test/speech-queue-regression.test.mjs test/testing.test.mjs test/test-sets.test.mjs",
     "check": "node tools/check-source.mjs && npm test",

package/runtime/arcane/components/theme-switcher.html

@@ -15,8 +15,8 @@
         max-inline-size:100%;
         padding:.2rem;
         border:1px solid var(--app-bar-border,var(--border-color,rgba(255,255,255,.25)));
-        border-radius:1.6rem;
-        background:color-mix(in srgb,currentColor 8%,transparent);
+        border-radius:var(--theme-switcher-radius,1.6rem);
+        background:var(--theme-switcher-background,color-mix(in srgb,currentColor 8%,transparent));
     }
 
     button {
@@ -28,11 +28,11 @@
         min-block-size:2.75rem;
         padding:.5rem .75rem;
         border:0;
-        border-radius:1.4rem;
+        border-radius:var(--theme-switcher-option-radius,1.4rem);
         background:transparent;
         color:inherit;
         cursor:pointer;
-        font:700 .875rem/1.25 system-ui,sans-serif;
+        font:var(--theme-switcher-font,700 .875rem/1.25 system-ui,sans-serif);
         text-align:center;
         white-space:normal;
         overflow-wrap:anywhere;

package/runtime/arcane/modules/AI.js

@@ -1182,7 +1182,6 @@ class AI {
     #speechJobSequence=0;
     #stopOllamaReady=null;
     #ttsSegmentation={...DEFAULT_TTS_SEGMENTATION};
-    #preparedSpeechReadiness = null;
     #preparedSpeechPlaybacks = new Set();
     #nativeSpeechPlaybacks = new Set();
     #speechUnlockContext = null;
@@ -4394,7 +4393,7 @@ class AI {
                     runtime.#browserSpeechTransition,activation.controller.signal
                 );
                 activation.controller.signal.throwIfAborted();
-                await runtime.#providerRuntime.setSpeechMuted(false);
+                await runtime.#providerRuntime.setSpeechMuted(false, {signal: activation.controller.signal});
                 if(runtime.#speechActivation!==activation)return false;
                 const status=runtime.#providerRuntime.status('tts');
                 runtime.muted=status.state!=='ready'||status.loaded!==true;
@@ -6409,22 +6408,7 @@ class AI {
             if(selected) {
                 const status = runtime.#providerRuntime.status('tts');
                 if(status.state !== 'ready' || status.loaded !== true) {
-                    if(!runtime.#preparedSpeechReadiness) {
-                        const readiness = runtime.#providerRuntime.setSpeechMuted(false);
-                        runtime.#preparedSpeechReadiness = readiness;
-                        readiness.finally(
-                            function releasePreparedSpeechReadiness() {
-                                if(runtime.#preparedSpeechReadiness === readiness) {
-                                    runtime.#preparedSpeechReadiness = null;
-                                }
-                            }
-                        ).catch(
-                            function reportPreparedSpeechReadinessFailure(error) {
-                                runtime.#traceSpeech('prepareTTS.readiness.error', {error});
-                            }
-                        );
-                    }
-                    await runtime.#preparedSpeechReadiness;
+                    await runtime.#providerRuntime.setSpeechMuted(false, {signal: requestSignal});
                     assertPreparedSpeechSelection();
                 }
             }

package/runtime/arcane/modules/AIProviderRuntime.js

@@ -1688,6 +1688,7 @@ export class AIProviderRuntime {
     #speechDesiredMuted = true;
     #speechTransition = Promise.resolve();
     #speechTransitionOwners = 0;
+    #speechActivation = null;
     #unsubscribeIntents = null;
     #closed = false;
     #closing = false;
@@ -2784,6 +2785,11 @@ export class AIProviderRuntime {
         }
         if (slot.loadPromise) {
             if (this.#sameSelection(slot.selection, targetSelection)) {
+                const activation = this.#speechActivation;
+                if (role === 'tts' && activation?.loadPromise === slot.loadPromise
+                    && signal !== activation.controller.signal) {
+                    return this.#observeSpeechActivation(activation, signal);
+                }
                 return slot.loadPromise;
             }
             return Promise.reject(
@@ -2957,6 +2963,10 @@ export class AIProviderRuntime {
             }
         );
         slot.loadPromise = loadOperation;
+        if (role === 'tts' && this.#speechActivation
+            && signal === this.#speechActivation.controller.signal) {
+            this.#speechActivation.loadPromise = loadOperation;
+        }
         publishAIRuntimeRoleState(
             role,
             roleRecord(
@@ -4270,19 +4280,34 @@ export class AIProviderRuntime {
         return true;
     }
 
-    async setSpeechMuted(muted) {
+    async setSpeechMuted(muted, {signal = null} = {}) {
         arcaneLogging.debug('[Arcane speech runtime] setSpeechMuted', muted);
         this.#assertOpen();
         this.#assertNotConfiguring();
         if (!is.boolean(muted)) {
             fail('AI speech muted state must be a boolean.');
         }
+        assertAbortSignal(signal);
+        if (signal?.aborted) throw normalizedAbort(signal.reason);
+        if (!muted && !this.#speechDesiredMuted && this.#speechActivation
+            && !this.#speechActivation.controller.signal.aborted) {
+            return this.#observeSpeechActivation(this.#speechActivation, signal);
+        }
         this.#speechDesiredMuted = muted;
         if (muted) {
             this.#speechMuted = true;
+            this.#speechActivation?.controller.abort();
             this.#slots.tts.loadController?.abort();
             this.cancel('tts');
         }
+        const activation = muted ? null : {
+            controller: new AbortController(),
+            interests: new Set(),
+            promise: null,
+            loadPromise: null,
+            settled: false
+        };
+        this.#speechActivation = activation;
         const runtime = this;
         this.#speechTransitionOwners += 1;
         const transition = this.#speechTransition.catch(
@@ -4290,6 +4315,9 @@ export class AIProviderRuntime {
         ).then(
             async function reconcileLatestAISpeechPreference() {
                 while (true) {
+                    if (activation?.controller.signal.aborted) {
+                        throw normalizedAbort(activation.controller.signal.reason);
+                    }
                     const desiredMuted = runtime.#speechDesiredMuted;
                     if (desiredMuted) {
                         runtime.cancel('tts');
@@ -4306,7 +4334,11 @@ export class AIProviderRuntime {
                         if (runtime.#speechDesiredMuted) {
                             continue;
                         }
-                        await runtime.load('tts');
+                        const currentActivation = runtime.#speechActivation;
+                        if (currentActivation?.controller.signal.aborted) {
+                            throw normalizedAbort(currentActivation.controller.signal.reason);
+                        }
+                        await runtime.load('tts', {signal: currentActivation?.controller.signal ?? null});
                     }
                     if (runtime.#speechDesiredMuted === desiredMuted) {
                         runtime.#speechMuted = desiredMuted;
@@ -4318,10 +4350,52 @@ export class AIProviderRuntime {
         ).finally(
             function releaseAIProviderSpeechTransitionOwnership() {
                 runtime.#speechTransitionOwners -= 1;
+                if (activation) {
+                    activation.settled = true;
+                    if (runtime.#speechActivation === activation) {
+                        runtime.#speechActivation = null;
+                    }
+                }
             }
         );
         this.#speechTransition = transition;
-        return transition;
+        if (!activation) return transition;
+        activation.promise = transition;
+        return this.#observeSpeechActivation(activation, signal);
+    }
+
+    #observeSpeechActivation(activation, signal) {
+        return new Promise(function observeAIProviderSpeechActivation(resolve, reject) {
+            const interest = {};
+            let settled = false;
+            activation.interests.add(interest);
+            function releaseSpeechActivationInterest() {
+                activation.interests.delete(interest);
+                signal?.removeEventListener('abort', cancelSpeechActivationInterest);
+            }
+            function cancelSpeechActivationInterest() {
+                if (settled) return;
+                settled = true;
+                releaseSpeechActivationInterest();
+                if (!activation.settled && activation.interests.size === 0) {
+                    activation.controller.abort(signal?.reason);
+                }
+                reject(normalizedAbort(signal?.reason));
+            }
+            signal?.addEventListener('abort', cancelSpeechActivationInterest, {once: true});
+            activation.promise.then(function speechActivationReady(status) {
+                if (settled) return;
+                settled = true;
+                releaseSpeechActivationInterest();
+                resolve(status);
+            }, function speechActivationFailed(error) {
+                if (settled) return;
+                settled = true;
+                releaseSpeechActivationInterest();
+                reject(error);
+            });
+            if (signal?.aborted) cancelSpeechActivationInterest();
+        });
     }
 
     #providerFor(slot) {