dsh-live-voice 0.0.2 → 0.1.0

package/README.md

@@ -100,7 +100,7 @@ Speech processing can run locally, but the DSH language model may still be remot
 
 ## Qwen3 HTTP engine
 
-When a compatible Qwen3 speech API is already running on the DSH host, choose **Qwen3 ASR — local MLX server** under Speech recognition and **Qwen3 TTS — local MLX server** under Speech output. Configure its loopback base URL in **DSH Settings → Live Voice**. The plugin supports the OminiX-API contract and standard OpenAI-style speech endpoints at `GET /health`, `POST /v1/audio/transcriptions`, and `POST /v1/audio/speech`; it does not install, start, stop, or manage that external service or its model weights.
+When a compatible Qwen3 speech API is already running on the DSH host, choose **Qwen3 ASR — local MLX server** under Speech recognition and **Qwen3 TTS — local MLX server** under Speech output. Configure its base URL in **DSH Settings → Live Voice**. The Qwen server may use any HTTP or HTTPS base URL reachable from the DSH host. The plugin supports the OminiX-API contract and standard OpenAI-style speech endpoints at `GET /health`, `POST /v1/audio/transcriptions`, and `POST /v1/audio/speech`; it does not install, start, stop, or manage that external service or its model weights.
 
 ## License
 

package/lib/client.js

@@ -435,8 +435,15 @@ var VoiceCoordinator = class {
       const expected = this.autoSendDraft;
       this.autoSendDraft = null;
       this.patch({ autoSendAt: null });
-      if (!this.disposed && this.snapshot.settings.sendingMode === "automatic" && expected === this.composer.getDraft())
-        this.composer.submit();
+      if (!this.disposed && this.snapshot.settings.sendingMode === "automatic" && expected === this.composer.getDraft()) {
+        try {
+          this.composer.submit();
+          this.transcript.reset();
+          this.recognition.reset?.();
+        } catch (error) {
+          this.patch({ error: message(error) });
+        }
+      }
     }, delay);
   }
   cancelAutoSend() {
@@ -1271,6 +1278,26 @@ var BrowserRecognitionEngine = class {
     };
     recognition.start();
   }
+  reset() {
+    const session = this.session;
+    const recognition = session?.recognition;
+    if (!session || !recognition) return false;
+    this._detach(recognition);
+    session.recognition = null;
+    try {
+      recognition.abort();
+    } catch {
+    }
+    this._activity(session, false);
+    if (this.session !== session || session.signal?.aborted) return false;
+    try {
+      this._begin(session);
+      return true;
+    } catch (error) {
+      this._fail(session, error);
+      return false;
+    }
+  }
   async stop() {
     const session = this.session;
     if (!session) return;
@@ -1854,7 +1881,7 @@ function createQwenSettings(React2) {
       h(
         "p",
         null,
-        "Host-wide settings for the resident Qwen3 ASR + TTS server. Only unauthenticated loopback HTTP is allowed; the browser reaches it through authenticated DSH routes."
+        "Host-wide settings for the Qwen3 ASR + TTS server. Enter any HTTP or HTTPS base URL reachable from the DSH host. The browser accesses it through authenticated DSH routes."
       ),
       field("Qwen API base URL", "baseUrl"),
       field("Request timeout (ms)", "timeoutMs", "number"),

package/lib/server.js

@@ -473,14 +473,14 @@ var defaultSettings = Object.freeze({
 });
 
 // src/engines/qwen-http-host.ts
-var LOOPBACK2 = /* @__PURE__ */ new Set(["localhost", "127.0.0.1", "::1", "[::1]"]);
 var clean2 = (value) => String(value ?? "").trim();
 function resolveQwenBaseUrl(value = process.env.DSH_LIVE_VOICE_QWEN_URL || "http://127.0.0.1:8080/") {
   const url = new URL(value);
-  if (url.protocol !== "http:" || !LOOPBACK2.has(url.hostname) || url.username || url.password || url.hash)
-    throw new Error("Qwen API URL must be an unauthenticated loopback http URL.");
+  if (!["http:", "https:"].includes(url.protocol))
+    throw new Error("Qwen API URL must use HTTP or HTTPS.");
   url.pathname = url.pathname.replace(/\/*$/, "/");
   url.search = "";
+  url.hash = "";
   return url;
 }
 function validateQwenConfig(value) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "dsh-live-voice",
-  "version": "0.0.2",
+  "version": "0.1.0",
   "type": "module",
   "scripts": {
     "typecheck": "tsc --noEmit",

package/src/client/qwen-settings.ts

@@ -113,7 +113,7 @@ export function createQwenSettings(React) {
       h(
         'p',
         null,
-        'Host-wide settings for the resident Qwen3 ASR + TTS server. Only unauthenticated loopback HTTP is allowed; the browser reaches it through authenticated DSH routes.',
+        'Host-wide settings for the Qwen3 ASR + TTS server. Enter any HTTP or HTTPS base URL reachable from the DSH host. The browser accesses it through authenticated DSH routes.',
       ),
       field('Qwen API base URL', 'baseUrl'),
       field('Request timeout (ms)', 'timeoutMs', 'number'),

package/src/core/coordinator.ts

@@ -317,8 +317,18 @@ export class VoiceCoordinator {
         !this.disposed &&
         this.snapshot.settings.sendingMode === 'automatic' &&
         expected === this.composer.getDraft()
-      )
-        this.composer.submit();
+      ) {
+        try {
+          this.composer.submit();
+          // Web Speech keeps a cumulative native result list for the lifetime of
+          // one recognition instance. Start a fresh instance at the turn boundary
+          // so the sent utterance cannot prefix the next one.
+          this.transcript.reset();
+          this.recognition.reset?.();
+        } catch (error) {
+          this.patch({ error: message(error) });
+        }
+      }
     }, delay);
   }
   cancelAutoSend() {

package/src/engines/qwen-http-host.ts

@@ -6,22 +6,16 @@ import { randomUUID } from 'node:crypto';
 import { defaultQwenVoice, isQwenVoice } from '../core/settings.ts';
 import { validateMonoPcm16Wav } from './recognition/whisper-http-host.ts';
 
-const LOOPBACK = new Set(['localhost', '127.0.0.1', '::1', '[::1]']);
 const clean = (value) => String(value ?? '').trim();
 export function resolveQwenBaseUrl(
   value = process.env.DSH_LIVE_VOICE_QWEN_URL || 'http://127.0.0.1:8080/',
 ) {
   const url = new URL(value);
-  if (
-    url.protocol !== 'http:' ||
-    !LOOPBACK.has(url.hostname) ||
-    url.username ||
-    url.password ||
-    url.hash
-  )
-    throw new Error('Qwen API URL must be an unauthenticated loopback http URL.');
+  if (!['http:', 'https:'].includes(url.protocol))
+    throw new Error('Qwen API URL must use HTTP or HTTPS.');
   url.pathname = url.pathname.replace(/\/*$/, '/');
   url.search = '';
+  url.hash = '';
   return url;
 }
 export function validateQwenConfig(value) {

package/src/engines/recognition/browser.ts

@@ -19,6 +19,7 @@ const notify = (callback, value) => {
  * not when permission is granted. Native asynchronous errors go to onError(Error).
  * onResult({interim,final}) contains the current interim string and NEW final text only.
  * onActivity(boolean) reports speech activity, not microphone permission/readiness.
+ * reset() replaces active native capture so its cumulative result list starts empty.
  * stop() aborts capture, discards pending results, removes handlers and cancels restarts.
  * Consecutive no-progress restarts are bounded; speech or nonempty results reset the budget.
  * Native start acceptance alone is not progress. Restart delay backs off up to 30 seconds.
@@ -272,6 +273,28 @@ export class BrowserRecognitionEngine {
     recognition.start();
   }
 
+  reset() {
+    const session = this.session;
+    const recognition = session?.recognition;
+    if (!session || !recognition) return false;
+    this._detach(recognition);
+    session.recognition = null;
+    try {
+      recognition.abort();
+    } catch {
+      /* Browser already ended capture. */
+    }
+    this._activity(session, false);
+    if (this.session !== session || session.signal?.aborted) return false;
+    try {
+      this._begin(session);
+      return true;
+    } catch (error) {
+      this._fail(session, error);
+      return false;
+    }
+  }
+
   async stop() {
     const session = this.session;
     if (!session) return;