arcane-os 0.11.1 → 0.11.3

package/CHANGELOG.md

@@ -1,5 +1,19 @@
 # Changelog
 
+## 0.11.3
+
+- Require Wllama's model-context load result to report success before publishing
+  model readiness. Failed initialization now follows the existing cleanup and
+  error path instead of allowing conversation requests against a failed context.
+  Preserve CPU and GPU selection, cached models, complete payloads, and cancellation.
+
+## 0.11.2
+
+- Show mobile speech feedback only during requested or active voice loading.
+  Align pending local-model activation to the top of the chat area at viewport
+  widths up to `44rem`. Preserve existing mobile controls, complete status and
+  error data, desktop layout, and model lifecycle.
+
 ## 0.11.1
 
 - Keep the shared Chat transcript, composer, speech controls, language selector,

package/README.md

@@ -19,7 +19,7 @@ version-locked SDK runtime, while an integrated Arcane checkout uses its live
 `arcane/` runtime. Both profiles preserve the same app URLs, theme, packaging,
 event, cancellation, and browser run contracts.
 
-This checkout defines the `0.11.1` SDK contract. Applications pin one exact npm
+This checkout defines the `0.11.3` SDK contract. Applications pin one exact npm
 version and lockfile; registry state is deliberately not baked into application
 artifacts.
 
@@ -41,7 +41,7 @@ Create one browser application, install its pinned SDK, and start its source
 server:
 
 ```bash
-npx arcane-os@0.11.1 new hello-speech --path ./hello-speech --target browser
+npx arcane-os@0.11.3 new hello-speech --path ./hello-speech --target browser
 cd hello-speech
 npm install
 npm run dev
@@ -392,7 +392,7 @@ uses the same controller for automatic memory extraction.
 Create a new repository-shaped Arcane application with the exact stable SDK:
 
 ```bash
-npx arcane-os@0.11.1 new my-app --path ./my-app --target portable --git
+npx arcane-os@0.11.3 new my-app --path ./my-app --target portable --git
 cd my-app
 npm install
 npm run dev
@@ -402,7 +402,7 @@ To enroll an existing repository, install the exact SDK and initialize only
 missing Arcane files:
 
 ```bash
-npm install --save-dev --save-exact arcane-os@0.11.1
+npm install --save-dev --save-exact arcane-os@0.11.3
 npm exec -- arcane init my-app --target portable
 ```
 
@@ -418,7 +418,7 @@ npm exec -- arcane-os targets
 No global SDK install or standalone Arcane CLI is required. The application
 repository's exact npm dependency and lockfile own the CLI and toolchain version.
 
-Use `npx arcane-os@0.11.1` for the initial bootstrap because it names this npm
+Use `npx arcane-os@0.11.3` for the initial bootstrap because it names this npm
 package explicitly; bare `npx arcane` outside an installed project could resolve
 a different package. Both installed commands invoke the same headless toolchain.
 Project-local npm scripts use the SDK pinned by that app's `package-lock.json`,
@@ -438,7 +438,7 @@ node ./bin/arcane.mjs new local-app --path ../local-app --target portable --git
 
 # From the generated app repository
 cd ../local-app
-npm install --save-dev --save-exact ../arcane-os-sdk/arcane-os-0.11.1.tgz
+npm install --save-dev --save-exact ../arcane-os-sdk/arcane-os-0.11.3.tgz
 npm ci
 ```
 
@@ -447,7 +447,7 @@ same location. The lockfile retains the selected package dependency while
 Arcane uses the installed package name and version. Local directory `file:` dependencies are not
 accepted because npm may install them as links; use a packed `.tgz`. A GitHub
 runner also needs that tarball at the locked path. After publication, replace
-the local declaration with the exact `arcane-os@0.11.1` registry package and
+the local declaration with the exact `arcane-os@0.11.3` registry package and
 commit the regenerated lock.
 
 Generated repositories use `npm ci --ignore-scripts` in CI. Run dependency
@@ -586,7 +586,7 @@ package installation, or assertions.
 
 ## Current target support
 
-Version `0.11.1` exposes one browser target and five explicitly paired
+Version `0.11.3` exposes one browser target and five explicitly paired
 native development targets: a non-runnable portable directory, a
 Windows x64 unsigned-local-test EXE bundle, Linux x64 and Linux ARM64
 unsigned-local-test DEBs, and an Android development-signed APK. The

package/browser-runtime/ai/browser-wllama-runtime.mjs

@@ -575,6 +575,13 @@ export function createPackagedWllamaRuntime({ logger = arcaneLogging } = {}) {
           "Wllama did not confirm a successfully loaded model.",
         );
       }
+        const loadedContext = next.getLoadedContextInfo();
+        if (loadedContext.success !== true) {
+            throw runtimeFailure(
+                'ARCANE_AI_LOAD_FAILED',
+                'Wllama failed to create the model context.'
+            );
+        }
       let adapter = null;
       if (webgpuRequired) {
           const telemetryOperation = trackOperation(

package/docs/reference/ai/browser-wasm.md

@@ -146,8 +146,11 @@ bar is indeterminate. Downloaded-file completion is not activation completion.
 The metadata tensor count and GPU layer assignment are descriptive only:
 upstream reports assigned layers before model-weight loading has finished.
 The pinned runtime does not expose a measured overall initialization fraction
-or a shader-compilation completion count. Readiness still requires the load
-operation to resolve and the runtime to confirm the model is loaded.
+or a shader-compilation completion count. Readiness requires the load operation
+to resolve, Wllama to report the model loaded, and its public
+`getLoadedContextInfo()` result to report `success:true`. A failed model context
+rejects loading with `ARCANE_AI_LOAD_FAILED`, releases the owned session, and
+leaves the provider in `error` with `loaded:false` before inference can begin.
 
 While a load remains active, the provider repeats its current record every five
 seconds with `heartbeat:true` and an updated `elapsedMs`. During initialization,

package/docs/reference/runtime-components.md

@@ -417,6 +417,8 @@ local LLM, the transcript, session status, composer, upload, speech panel
 (including application-supplied controls), language selector, and timer stay
 hidden until that exact sticky role is `ready` and `loaded`. The existing model
 activation area remains available for Start, progress, cancellation, and retry.
+At viewport widths up to `44rem`, that pending local-model activation area sits
+at the top of the chat area; wider layouts retain their existing placement.
 A selected cloud route (`localOnly:false`) shows the conversation surfaces
 immediately; Send still follows actual LLM availability. Later unload, failure,
 or selection changes update visibility from the same runtime subscription
@@ -1180,8 +1182,12 @@ selects a Speech interface for voice output. Transcription activation,
 recording, sending recorded audio, and their status/progress are unavailable;
 `requestSTTActivation()` returns `false` without requesting a provider action.
 The keyboard remains the voice-input path. The existing TTS mute/unmute control
-and voice status remain available, and Chat retains its End control. Desktop
-speech and the dedicated `voice-transcription.html` component retain their
+remains available alongside application-supplied controls such as End. Mobile
+voice feedback participates in layout only while a voice load is requested or
+the TTS role is `loading`; it collapses in other states, including unloaded, ready, and
+speaking. Complete status text and error events remain maintained by the same
+owner, and the voice control retains its state-specific labels and retry action.
+Desktop speech and the dedicated `voice-transcription.html` component retain their
 existing behavior.
 
 On desktop, the Hold to talk control remains disabled unless sticky STT state

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "arcane-os",
-  "version": "0.11.1",
+  "version": "0.11.3",
   "description": "Arcane OS JavaScript SDK, project-local CLI, browser runtime, and repository-portable application packager.",
   "type": "module",
   "main": "./src/index.mjs",

package/runtime/arcane/components/chat.html

@@ -531,6 +531,10 @@
             padding: var(--arcane-space-2, .5rem);
         }
 
+        .chat_area[data-model-start-required="true"]:not([data-conversation-complete="true"]) {
+            justify-content: flex-start;
+        }
+
         .chat_output > li {
             max-width: 96%;
         }

package/runtime/arcane/components/speech.html

@@ -147,6 +147,10 @@
         grid-column: 1;
     }
 
+    :host([data-device-class="mobile"]:not([data-voice-loading="true"])) .speech_feedback {
+        display: none;
+    }
+
     :host([data-device-class="mobile"]) #muteButton {
         grid-row: 1;
         grid-column: 2;
@@ -725,6 +729,9 @@
     }
 
     function renderStatus() {
+        host.dataset.voiceLoading = String(
+            pendingTTSIntent === 'load' || ttsRole.state === 'loading'
+        );
         const messages = [];
         let tone = '';