voice-router-dev 0.1.8 → 0.2.0

package/dist/index.mjs

@@ -210,6 +210,69 @@ function createVoiceRouter(config, adapters) {
   return router;
 }
 
+// src/generated/deepgram/schema/listenV1EncodingParameter.ts
+var ListenV1EncodingParameter = {
+  linear16: "linear16",
+  flac: "flac",
+  mulaw: "mulaw",
+  opus: "opus",
+  speex: "speex",
+  g729: "g729"
+};
+
+// src/generated/deepgram/schema/speakV1EncodingParameter.ts
+var SpeakV1EncodingParameter = {
+  linear16: "linear16",
+  aac: "aac",
+  opus: "opus",
+  mp3: "mp3",
+  flac: "flac",
+  mulaw: "mulaw",
+  alaw: "alaw"
+};
+
+// src/generated/deepgram/schema/speakV1ContainerParameter.ts
+var SpeakV1ContainerParameter = {
+  none: "none",
+  wav: "wav",
+  ogg: "ogg"
+};
+
+// src/generated/deepgram/schema/speakV1SampleRateParameter.ts
+var SpeakV1SampleRateParameter = {
+  NUMBER_16000: 16e3,
+  NUMBER_24000: 24e3,
+  NUMBER_32000: 32e3,
+  NUMBER_48000: 48e3,
+  null: null,
+  NUMBER_8000: 8e3,
+  NUMBER_22050: 22050
+};
+
+// src/generated/gladia/schema/streamingSupportedEncodingEnum.ts
+var StreamingSupportedEncodingEnum = {
+  "wav/pcm": "wav/pcm",
+  "wav/alaw": "wav/alaw",
+  "wav/ulaw": "wav/ulaw"
+};
+
+// src/generated/gladia/schema/streamingSupportedSampleRateEnum.ts
+var StreamingSupportedSampleRateEnum = {
+  NUMBER_8000: 8e3,
+  NUMBER_16000: 16e3,
+  NUMBER_32000: 32e3,
+  NUMBER_44100: 44100,
+  NUMBER_48000: 48e3
+};
+
+// src/generated/gladia/schema/streamingSupportedBitDepthEnum.ts
+var StreamingSupportedBitDepthEnum = {
+  NUMBER_8: 8,
+  NUMBER_16: 16,
+  NUMBER_24: 24,
+  NUMBER_32: 32
+};
+
 // src/constants/defaults.ts
 var DEFAULT_TIMEOUTS = {
   /** Standard HTTP request timeout for API calls (60 seconds) */
@@ -1097,21 +1160,6 @@ var StreamingResponseStatus = {
   error: "error"
 };
 
-// src/generated/gladia/schema/streamingSupportedBitDepthEnum.ts
-var StreamingSupportedBitDepthEnum = {
-  NUMBER_8: 8,
-  NUMBER_16: 16,
-  NUMBER_24: 24,
-  NUMBER_32: 32
-};
-
-// src/generated/gladia/schema/streamingSupportedEncodingEnum.ts
-var StreamingSupportedEncodingEnum = {
-  "wav/pcm": "wav/pcm",
-  "wav/alaw": "wav/alaw",
-  "wav/ulaw": "wav/ulaw"
-};
-
 // src/generated/gladia/schema/streamingSupportedModels.ts
 var StreamingSupportedModels = {
   "solaria-1": "solaria-1"
@@ -1123,15 +1171,6 @@ var StreamingSupportedRegions = {
   "eu-west": "eu-west"
 };
 
-// src/generated/gladia/schema/streamingSupportedSampleRateEnum.ts
-var StreamingSupportedSampleRateEnum = {
-  NUMBER_8000: 8e3,
-  NUMBER_16000: 16e3,
-  NUMBER_32000: 32e3,
-  NUMBER_44100: 44100,
-  NUMBER_48000: 48e3
-};
-
 // src/generated/gladia/schema/subtitlesFormatEnum.ts
 var SubtitlesFormatEnum = {
   srt: "srt",
@@ -2434,9 +2473,6 @@ var createTranscript = (transcriptParams, options) => {
 var getTranscript = (transcriptId, options) => {
   return axios2.get(`/v2/transcript/${transcriptId}`, options);
 };
-var createTemporaryToken = (createRealtimeTemporaryTokenParams, options) => {
-  return axios2.post("/v2/realtime/token", createRealtimeTemporaryTokenParams, options);
-};
 
 // src/adapters/assemblyai-adapter.ts
 var AssemblyAIAdapter = class extends BaseAdapter {
@@ -2456,8 +2492,9 @@ var AssemblyAIAdapter = class extends BaseAdapter {
     };
     this.baseUrl = "https://api.assemblyai.com";
     // Generated functions already include /v2 path
-    this.wsBaseUrl = "wss://api.assemblyai.com/v2/realtime/ws";
+    this.wsBaseUrl = "wss://streaming.assemblyai.com/v3/ws";
   }
+  // v3 Universal Streaming endpoint
   /**
    * Get axios config for generated API client functions
    * Configures headers and base URL using authorization header
@@ -2781,14 +2818,17 @@ var AssemblyAIAdapter = class extends BaseAdapter {
    */
   async transcribeStream(options, callbacks) {
     this.validateConfig();
-    const tokenResponse = await createTemporaryToken(
-      { expires_in: 3600 },
-      // Token expires in 1 hour
-      this.getAxiosConfig()
-    );
-    const token = tokenResponse.data.token;
-    const wsUrl = `${this.wsBaseUrl}?sample_rate=${options?.sampleRate || 16e3}&token=${token}`;
-    const ws = new WebSocket2(wsUrl);
+    if (!this.config?.apiKey) {
+      throw new Error("API key is required for streaming");
+    }
+    const sampleRate = options?.sampleRate || 16e3;
+    const encoding = options?.encoding ? mapEncodingToProvider(options.encoding, "assemblyai") : "pcm_s16le";
+    const wsUrl = `${this.wsBaseUrl}?sample_rate=${sampleRate}&encoding=${encoding}`;
+    const ws = new WebSocket2(wsUrl, {
+      headers: {
+        Authorization: this.config.apiKey
+      }
+    });
     let sessionStatus = "connecting";
     const sessionId = `assemblyai-${Date.now()}-${Math.random().toString(36).substring(7)}`;
     ws.on("open", () => {
@@ -2798,41 +2838,42 @@ var AssemblyAIAdapter = class extends BaseAdapter {
     ws.on("message", (data) => {
       try {
         const message = JSON.parse(data.toString());
-        if (message.message_type === "SessionBegins") {
+        if ("error" in message) {
+          callbacks?.onError?.({
+            code: "API_ERROR",
+            message: message.error
+          });
+          return;
+        }
+        if (message.type === "Begin") {
+          const beginMsg = message;
           callbacks?.onMetadata?.({
-            sessionId: message.session_id,
-            expiresAt: message.expires_at
+            sessionId: beginMsg.id,
+            expiresAt: new Date(beginMsg.expires_at).toISOString()
           });
-        } else if (message.message_type === "PartialTranscript") {
+        } else if (message.type === "Turn") {
+          const turnMsg = message;
           callbacks?.onTranscript?.({
             type: "transcript",
-            text: message.text,
-            isFinal: false,
-            confidence: message.confidence,
-            words: message.words.map((word) => ({
+            text: turnMsg.transcript,
+            isFinal: turnMsg.end_of_turn,
+            confidence: turnMsg.end_of_turn_confidence,
+            words: turnMsg.words.map((word) => ({
               text: word.text,
               start: word.start / 1e3,
+              // Convert ms to seconds
               end: word.end / 1e3,
               confidence: word.confidence
             })),
-            data: message
+            data: turnMsg
           });
-        } else if (message.message_type === "FinalTranscript") {
-          callbacks?.onTranscript?.({
-            type: "transcript",
-            text: message.text,
-            isFinal: true,
-            confidence: message.confidence,
-            words: message.words.map((word) => ({
-              text: word.text,
-              start: word.start / 1e3,
-              end: word.end / 1e3,
-              confidence: word.confidence
-            })),
-            data: message
+        } else if (message.type === "Termination") {
+          const termMsg = message;
+          callbacks?.onMetadata?.({
+            terminated: true,
+            audioDurationSeconds: termMsg.audio_duration_seconds,
+            sessionDurationSeconds: termMsg.session_duration_seconds
           });
-        } else if (message.message_type === "SessionTerminated") {
-          callbacks?.onMetadata?.({ terminated: true });
         }
       } catch (error) {
         callbacks?.onError?.({
@@ -5071,9 +5112,16 @@ export {
   GladiaAdapter,
   schema_exports as GladiaTypes,
   GladiaWebhookHandler,
+  ListenV1EncodingParameter,
   OpenAIWhisperAdapter,
+  SpeakV1ContainerParameter,
+  SpeakV1EncodingParameter,
+  SpeakV1SampleRateParameter,
   SpeechmaticsAdapter,
   SpeechmaticsWebhookHandler,
+  StreamingSupportedBitDepthEnum,
+  StreamingSupportedEncodingEnum,
+  StreamingSupportedSampleRateEnum,
   VoiceRouter,
   WebhookRouter,
   createAssemblyAIAdapter,