@mastra/voice-deepgram 0.0.0-mastra-3338-mastra-memory-pinecone-20250507174328 → 0.0.0-mastra-auto-detect-server-20260108233416

package/dist/docs/voice/02-reference.md

@@ -0,0 +1,62 @@
+# Voice API Reference
+
+> API reference for voice - 1 entries
+
+
+---
+
+## Reference: Deepgram
+
+> Documentation for the Deepgram voice implementation, providing text-to-speech and speech-to-text capabilities with multiple voice models and languages.
+
+The Deepgram voice implementation in Mastra provides text-to-speech (TTS) and speech-to-text (STT) capabilities using Deepgram's API. It supports multiple voice models and languages, with configurable options for both speech synthesis and transcription.
+
+## Usage Example
+
+```typescript
+import { DeepgramVoice } from "@mastra/voice-deepgram";
+
+// Initialize with default configuration (uses DEEPGRAM_API_KEY environment variable)
+const voice = new DeepgramVoice();
+
+// Initialize with custom configuration
+const voice = new DeepgramVoice({
+  speechModel: {
+    name: "aura",
+    apiKey: "your-api-key",
+  },
+  listeningModel: {
+    name: "nova-2",
+    apiKey: "your-api-key",
+  },
+  speaker: "asteria-en",
+});
+
+// Text-to-Speech
+const audioStream = await voice.speak("Hello, world!");
+
+// Speech-to-Text
+const transcript = await voice.listen(audioStream);
+```
+
+## Constructor Parameters
+
+### DeepgramVoiceConfig
+
+## Methods
+
+### speak()
+
+Converts text to speech using the configured speech model and voice.
+
+Returns: `Promise<NodeJS.ReadableStream>`
+
+### listen()
+
+Converts speech to text using the configured listening model.
+
+Returns: `Promise<string>`
+
+### getSpeakers()
+
+Returns a list of available voice options.

package/dist/index.cjs

@@ -26,6 +26,9 @@ var DEEPGRAM_VOICES = [
 var DeepgramVoice = class extends voice.MastraVoice {
   speechClient;
   listeningClient;
+  storedSpeechModel;
+  storedListeningModel;
+  storedSpeaker;
   constructor({
     speechModel,
     listeningModel,
@@ -51,6 +54,14 @@ var DeepgramVoice = class extends voice.MastraVoice {
       },
       speaker
     });
+    this.storedSpeechModel = {
+      name: speechModel?.name ?? defaultSpeechModel.name,
+      apiKey: speechModel?.apiKey ?? defaultSpeechModel.apiKey
+    };
+    this.storedListeningModel = {
+      name: listeningModel?.name ?? defaultListeningModel.name,
+      apiKey: listeningModel?.apiKey ?? defaultListeningModel.apiKey
+    };
     const speechApiKey = speechModel?.apiKey || defaultApiKey;
     const listeningApiKey = listeningModel?.apiKey || defaultApiKey;
     if (!speechApiKey && !listeningApiKey) {
@@ -62,14 +73,12 @@ var DeepgramVoice = class extends voice.MastraVoice {
     if (listeningApiKey) {
       this.listeningClient = sdk.createClient(listeningApiKey);
     }
-    this.speaker = speaker || "asteria-en";
+    this.storedSpeaker = speaker || "asteria-en";
   }
   async getSpeakers() {
-    return this.traced(async () => {
-      return DEEPGRAM_VOICES.map((voice) => ({
-        voiceId: voice
-      }));
-    }, "voice.deepgram.getSpeakers")();
+    return DEEPGRAM_VOICES.map((voice) => ({
+      voiceId: voice
+    }));
   }
   async speak(input, options) {
     if (!this.speechClient) {
@@ -92,49 +101,60 @@ var DeepgramVoice = class extends voice.MastraVoice {
     if (text.trim().length === 0) {
       throw new Error("Input text is empty");
     }
-    return this.traced(async () => {
-      if (!this.speechClient) {
-        throw new Error("No speech client configured");
-      }
-      let model;
-      if (options?.speaker) {
-        model = this.speechModel?.name + "-" + options.speaker;
-      } else if (this.speaker) {
-        model = this.speechModel?.name + "-" + this.speaker;
-      }
-      const speakClient = this.speechClient.speak;
-      const response = await speakClient.request(
-        { text },
-        {
-          model,
-          ...options
-        }
-      );
-      const webStream = await response.getStream();
-      if (!webStream) {
-        throw new Error("No stream returned from Deepgram");
+    const baseModel = this.storedSpeechModel?.name;
+    const speakerId = options?.speaker || this.storedSpeaker;
+    const modelName = baseModel && speakerId ? speakerId.startsWith(`${baseModel}-`) ? speakerId : `${baseModel}-${speakerId}` : baseModel || speakerId;
+    const speakClient = this.speechClient.speak;
+    const response = await speakClient.request(
+      { text },
+      {
+        model: modelName,
+        ...Object.fromEntries(Object.entries(options ?? {}).filter(([k]) => k !== "speaker"))
       }
-      const reader = webStream.getReader();
-      const nodeStream = new stream.PassThrough();
-      (async () => {
-        try {
-          while (true) {
-            const { done, value } = await reader.read();
-            if (done) {
-              nodeStream.end();
-              break;
-            }
-            nodeStream.write(value);
+    );
+    const webStream = await response.getStream();
+    if (!webStream) {
+      throw new Error("No stream returned from Deepgram");
+    }
+    const reader = webStream.getReader();
+    const nodeStream = new stream.PassThrough();
+    (async () => {
+      try {
+        while (true) {
+          const { done, value } = await reader.read();
+          if (done) {
+            nodeStream.end();
+            break;
           }
-        } catch (error) {
-          nodeStream.destroy(error);
+          nodeStream.write(value);
         }
-      })().catch((error) => {
+      } catch (error) {
         nodeStream.destroy(error);
-      });
-      return nodeStream;
-    }, "voice.deepgram.speak")();
+      }
+    })().catch((error) => {
+      nodeStream.destroy(error);
+    });
+    return nodeStream;
+  }
+  /**
+   * Checks if listening capabilities are enabled.
+   *
+   * @returns {Promise<{ enabled: boolean }>}
+   */
+  async getListener() {
+    return { enabled: true };
   }
+  /**
+   * Transcribes audio with optional speaker diarization.
+   *
+   * @param audioStream - Audio input stream
+   * @param options - Transcription options (diarize, language, etc.)
+   * @returns Promise resolving to:
+   *   - transcript: Full transcript string
+   *   - words: Array of word objects with timing and confidence
+   *   - raw: Complete Deepgram API response
+   *   - speakerSegments: (when diarize=true) Array of {word, speaker, start, end}
+   */
   async listen(audioStream, options) {
     if (!this.listeningClient) {
       throw new Error("Deepgram listening client not configured");
@@ -148,24 +168,41 @@ var DeepgramVoice = class extends voice.MastraVoice {
       }
     }
     const buffer = Buffer.concat(chunks);
-    return this.traced(async () => {
-      if (!this.listeningClient) {
-        throw new Error("No listening client configured");
-      }
-      const { result, error } = await this.listeningClient.listen.prerecorded.transcribeFile(buffer, {
-        model: this.listeningModel?.name,
-        ...options
-      });
-      if (error) {
-        throw error;
-      }
-      const transcript = result.results?.channels?.[0]?.alternatives?.[0]?.transcript;
-      if (!transcript) {
-        throw new Error("No transcript found in Deepgram response");
-      }
-      return transcript;
-    }, "voice.deepgram.listen")();
+    const { diarize, diarize_speaker_count: _, ...restOptions } = options ?? {};
+    const { result, error } = await this.listeningClient.listen.prerecorded.transcribeFile(buffer, {
+      ...restOptions,
+      model: this.storedListeningModel?.name,
+      diarize
+    });
+    if (error) {
+      throw error;
+    }
+    const channel = result.results?.channels?.[0];
+    const alt = channel?.alternatives?.[0];
+    if (!alt) {
+      return {
+        transcript: "",
+        words: [],
+        raw: result
+      };
+    }
+    const response = {
+      transcript: alt.transcript,
+      words: alt.words,
+      raw: result
+    };
+    if (diarize && alt.words) {
+      response.speakerSegments = alt.words.map((w) => ({
+        word: w.word,
+        speaker: w.speaker,
+        start: w.start,
+        end: w.end
+      }));
+    }
+    return response;
   }
 };
 
 exports.DeepgramVoice = DeepgramVoice;
+//# sourceMappingURL=index.cjs.map
+//# sourceMappingURL=index.cjs.map

package/dist/index.cjs.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/voices.ts","../src/index.ts"],"names":["MastraVoice","createClient","PassThrough"],"mappings":";;;;;;;;;AAKO,IAAM,eAAA,GAAkB;AAAA,EAC7B,YAAA;AAAA,EACA,SAAA;AAAA,EACA,WAAA;AAAA,EACA,WAAA;AAAA,EACA,SAAA;AAAA,EACA,UAAA;AAAA,EACA,UAAA;AAAA,EACA,YAAA;AAAA,EACA,UAAA;AAAA,EACA,YAAA;AAAA,EACA,WAAA;AAAA,EACA;AACF,CAAA;;;ACKO,IAAM,aAAA,GAAN,cAA4BA,iBAAA,CAAY;AAAA,EACrC,YAAA;AAAA,EACA,eAAA;AAAA,EACA,iBAAA;AAAA,EACA,oBAAA;AAAA,EACA,aAAA;AAAA,EAER,WAAA,CAAY;AAAA,IACV,WAAA;AAAA,IACA,cAAA;AAAA,IACA;AAAA,GACF,GAA4G,EAAC,EAAG;AAC9G,IAAA,MAAM,aAAA,GAAgB,QAAQ,GAAA,CAAI,gBAAA;AAElC,IAAA,MAAM,kBAAA,GAA+D;AAAA,MACnE,IAAA,EAAM,MAAA;AAAA,MACN,MAAA,EAAQ;AAAA,KACV;AAEA,IAAA,MAAM,qBAAA,GAAkE;AAAA,MACtE,IAAA,EAAM,MAAA;AAAA,MACN,MAAA,EAAQ;AAAA,KACV;AAEA,IAAA,KAAA,CAAM;AAAA,MACJ,WAAA,EAAa;AAAA,QACX,IAAA,EAAM,WAAA,EAAa,IAAA,IAAQ,kBAAA,CAAmB,IAAA;AAAA,QAC9C,MAAA,EAAQ,WAAA,EAAa,MAAA,IAAU,kBAAA,CAAmB;AAAA,OACpD;AAAA,MACA,cAAA,EAAgB;AAAA,QACd,IAAA,EAAM,cAAA,EAAgB,IAAA,IAAQ,qBAAA,CAAsB,IAAA;AAAA,QACpD,MAAA,EAAQ,cAAA,EAAgB,MAAA,IAAU,qBAAA,CAAsB;AAAA,OAC1D;AAAA,MACA;AAAA,KACD,CAAA;AAED,IAAA,IAAA,CAAK,iBAAA,GAAoB;AAAA,MACvB,IAAA,EAAM,WAAA,EAAa,IAAA,IAAQ,kBAAA,CAAmB,IAAA;AAAA,MAC9C,MAAA,EAAQ,WAAA,EAAa,MAAA,IAAU,kBAAA,CAAmB;AAAA,KACpD;AACA,IAAA,IAAA,CAAK,oBAAA,GAAuB;AAAA,MAC1B,IAAA,EAAM,cAAA,EAAgB,IAAA,IAAQ,qBAAA,CAAsB,IAAA;AAAA,MACpD,MAAA,EAAQ,cAAA,EAAgB,MAAA,IAAU,qBAAA,CAAsB;AAAA,KAC1D;AAEA,IAAA,MAAM,YAAA,GAAe,aAAa,MAAA,IAAU,aAAA;AAC5C,IAAA,MAAM,eAAA,GAAkB,gBAAgB,MAAA,IAAU,aAAA;AAElD,IAAA,IAAI,CAAC,YAAA,IAAgB,CAAC,eAAA,EAAiB;AACrC,MAAA,MAAM,IAAI,MAAM,4FAA4F,CAAA;AAAA,IAC9G;AAEA,IAAA,IAAI,YAAA,EAAc;AAChB,MAAA,IAAA,CAAK,YAAA,GAAeC,iBAAa,YAAY,CAAA;AAAA,IAC/C;AACA,IAAA,IAAI,eAAA,EAAiB;AACnB,MAAA,IAAA,CAAK,eAAA,GAAkBA,iBAAa,eAAe,CAAA;AAAA,IACrD;AAEA,IAAA,IAAA,CAAK,gBAAgB,OAAA,IAAW,YAAA;AAAA,EAClC;AAAA,EAEA,MAAM,WAAA,GAAc;AAClB,IAAA,OAAO,eAAA,CAAgB,IAAI,CAAA,KAAA,MAAU;AAAA,MACnC,OAAA,EAAS;AAAA,KACX,CAAE,CAAA;AAAA,EACJ;AAAA,EAEA,MAAM,KAAA,CACJ,KAAA,EACA,OAAA,EAIgC;AAChC,IAAA,IAAI,CAAC,KAAK,YAAA,EAAc;AACtB,MAAA,MAAM,IAAI,MAAM,uCAAuC,CAAA;AAAA,IACzD;AAEA,IAAA,IAAI,IAAA;AACJ,IAAA,IAAI,OAAO,UAAU,QAAA,EAAU;AAC7B,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,WAAA,MAAiB,SAAS,KAAA,EAAO;AAC/B,QAAA,IAAI,OAAO,UAAU,QAAA,EAAU;AAC7B,UAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,KAAK,CAAC,CAAA;AAAA,QAChC,CAAA,MAAO;AACL,UAAA,MAAA,CAAO,KAAK,KAAK,CAAA;AAAA,QACnB;AAAA,MACF;AACA,MAAA,IAAA,GAAO,MAAA,CAAO,MAAA,CAAO,MAAM,CAAA,CAAE,SAAS,OAAO,CAAA;AAAA,IAC/C,CAAA,MAAO;AACL,MAAA,IAAA,GAAO,KAAA;AAAA,IACT;AAEA,IAAA,IAAI,IAAA,CAAK,IAAA,EAAK,CAAE,MAAA,KAAW,CAAA,EAAG;AAC5B,MAAA,MAAM,IAAI,MAAM,qBAAqB,CAAA;AAAA,IACvC;AAEA,IAAA,MAAM,SAAA,GAAY,KAAK,iBAAA,EAAmB,IAAA;AAC1C,IAAA,MAAM,SAAA,GAAY,OAAA,EAAS,OAAA,IAAW,IAAA,CAAK,aAAA;AAE3C,IAAA,MAAM,SAAA,GACJ,SAAA,IAAa,SAAA,GACT,SAAA,CAAU,WAAW,CAAA,EAAG,SAAS,CAAA,CAAA,CAAG,CAAA,GAClC,YACA,CAAA,EAAG,SAAS,CAAA,CAAA,EAAI,SAAS,KAC3B,SAAA,IAAa,SAAA;AAEnB,IAAA,MAAM,WAAA,GAAc,KAAK,YAAA,CAAa,KAAA;AACtC,IAAA,MAAM,QAAA,GAAW,MAAM,WAAA,CAAY,OAAA;AAAA,MACjC,EAAE,IAAA,EAAK;AAAA,MACP;AAAA,QACE,KAAA,EAAO,SAAA;AAAA,QACP,GAAG,MAAA,CAAO,WAAA,CAAY,MAAA,CAAO,OAAA,CAAQ,WAAW,EAAE,CAAA,CAAE,MAAA,CAAO,CAAC,CAAC,CAAC,CAAA,KAAM,CAAA,KAAM,SAAS,CAAC;AAAA;AACtF,KACF;AAEA,IAAA,MAAM,SAAA,GAAY,MAAM,QAAA,CAAS,SAAA,EAAU;AAC3C,IAAA,IAAI,CAAC,SAAA,EAAW;AACd,MAAA,MAAM,IAAI,MAAM,kCAAkC,CAAA;AAAA,IACpD;AAEA,IAAA,MAAM,MAAA,GAAS,UAAU,SAAA,EAAU;AACnC,IAAA,MAAM,UAAA,GAAa,IAAIC,kBAAA,EAAY;AAGnC,IAAA,CAAC,YAAY;AACX,MAAA,IAAI;AACF,QAAA,OAAO,IAAA,EAAM;AACX,UAAA,MAAM,EAAE,IAAA,EAAM,KAAA,EAAM,GAAI,MAAM,OAAO,IAAA,EAAK;AAC1C,UAAA,IAAI,IAAA,EAAM;AACR,YAAA,UAAA,CAAW,GAAA,EAAI;AACf,YAAA;AAAA,UACF;AACA,UAAA,UAAA,CAAW,MAAM,KAAK,CAAA;AAAA,QACxB;AAAA,MACF,SAAS,KAAA,EAAO;AACd,QAAA,UAAA,CAAW,QAAQ,KAAc,CAAA;AAAA,MACnC;AAAA,IACF,CAAA,GAAG,CAAE,KAAA,CAAM,CAAA,KAAA,KAAS;AAClB,MAAA,UAAA,CAAW,QAAQ,KAAc,CAAA;AAAA,IACnC,CAAC,CAAA;AAED,IAAA,OAAO,UAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,MAAM,WAAA,GAAc;AAClB,IAAA,OAAO,EAAE,SAAS,IAAA,EAAK;AAAA,EACzB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,MAAM,MAAA,CACJ,WAAA,EACA,OAAA,EAIc;AACd,IAAA,IAAI,CAAC,KAAK,eAAA,EAAiB;AACzB,MAAA,MAAM,IAAI,MAAM,0CAA0C,CAAA;AAAA,IAC5D;AAEA,IAAA,MAAM,SAAmB,EAAC;AAC1B,IAAA,WAAA,MAAiB,SAAS,WAAA,EAAa;AACrC,MAAA,IAAI,OAAO,UAAU,QAAA,EAAU;AAC7B,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,KAAK,CAAC,CAAA;AAAA,MAChC,CAAA,MAAO;AACL,QAAA,MAAA,CAAO,KAAK,KAAK,CAAA;AAAA,MACnB;AAAA,IACF;AACA,IAAA,MAAM,MAAA,GAAS,MAAA,CAAO,MAAA,CAAO,MAAM,CAAA;AACnC,IAAA,MAAM,EAAE,SAAS,qBAAA,EAAuB,CAAA,EAAG,GAAG,WAAA,EAAY,GAAI,WAAW,EAAC;AAC1E,IAAA,MAAM,EAAE,MAAA,EAAQ,KAAA,EAAM,GAAI,MAAM,KAAK,eAAA,CAAgB,MAAA,CAAO,WAAA,CAAY,cAAA,CAAe,MAAA,EAAQ;AAAA,MAC7F,GAAG,WAAA;AAAA,MACH,KAAA,EAAO,KAAK,oBAAA,EAAsB,IAAA;AAAA,MAClC;AAAA,KACD,CAAA;AACD,IAAA,IAAI,KAAA,EAAO;AACT,MAAA,MAAM,KAAA;AAAA,IACR;AAEA,IAAA,MAAM,OAAA,GAAU,MAAA,CAAO,OAAA,EAAS,QAAA,GAAW,CAAC,CAAA;AAC5C,IAAA,MAAM,GAAA,GAKU,OAAA,EAAS,YAAA,GAAe,CAAC,CAAA;AAEzC,IAAA,IAAI,CAAC,GAAA,EAAK;AACR,MAAA,OAAO;AAAA,QACL,UAAA,EAAY,EAAA;AAAA,QACZ,OAAO,EAAC;AAAA,QACR,GAAA,EAAK;AAAA,OACP;AAAA,IACF;AAEA,IAAA,MAAM,QAAA,GAAgB;AAAA,MACpB,YAAY,GAAA,CAAI,UAAA;AAAA,MAChB,OAAO,GAAA,CAAI,KAAA;AAAA,MACX,GAAA,EAAK;AAAA,KACP;AAEA,IAAA,IAAI,OAAA,IAAW,IAAI,KAAA,EAAO;AACxB,MAAA,QAAA,CAAS,eAAA,GAAkB,GAAA,CAAI,KAAA,CAAM,GAAA,CAAI,CAAC,CAAA,MAAqB;AAAA,QAC7D,MAAM,CAAA,CAAE,IAAA;AAAA,QACR,SAAS,CAAA,CAAE,OAAA;AAAA,QACX,OAAO,CAAA,CAAE,KAAA;AAAA,QACT,KAAK,CAAA,CAAE;AAAA,OACT,CAAE,CAAA;AAAA,IACJ;AAEA,IAAA,OAAO,QAAA;AAAA,EACT;AACF","file":"index.cjs","sourcesContent":["/**\n * List of available Deepgram voice models for text-to-speech\n * Each voice is designed for specific use cases and languages\n * Format: {name}-{language} (e.g. asteria-en)\n */\nexport const DEEPGRAM_VOICES = [\n  'asteria-en',\n  'luna-en',\n  'stella-en',\n  'athena-en',\n  'hera-en',\n  'orion-en',\n  'arcas-en',\n  'perseus-en',\n  'angus-en',\n  'orpheus-en',\n  'helios-en',\n  'zeus-en',\n] as const;\n\nexport type DeepgramVoiceId = (typeof DEEPGRAM_VOICES)[number];\n\n/**\n * List of available Deepgram models for text-to-speech and speech-to-text\n */\nexport const DEEPGRAM_MODELS = ['aura', 'whisper', 'base', 'enhanced', 'nova', 'nova-2', 'nova-3'] as const;\n\nexport type DeepgramModel = (typeof DEEPGRAM_MODELS)[number];\n","import { PassThrough } from 'node:stream';\n\nimport { createClient } from '@deepgram/sdk';\nimport { MastraVoice } from '@mastra/core/voice';\n\nimport { DEEPGRAM_VOICES } from './voices';\nimport type { DeepgramVoiceId, DeepgramModel } from './voices';\n\ninterface DeepgramVoiceConfig {\n  name?: DeepgramModel;\n  apiKey?: string;\n  properties?: Record<string, any>;\n  language?: string;\n}\n\ninterface DeepgramWord {\n  word: string;\n  start?: number;\n  end?: number;\n  confidence?: number;\n  speaker?: number;\n}\n\nexport class DeepgramVoice extends MastraVoice {\n  private speechClient?: ReturnType<typeof createClient>;\n  private listeningClient?: ReturnType<typeof createClient>;\n  private storedSpeechModel?: { name: DeepgramModel; apiKey?: string };\n  private storedListeningModel?: { name: DeepgramModel; apiKey?: string };\n  private storedSpeaker?: DeepgramVoiceId;\n\n  constructor({\n    speechModel,\n    listeningModel,\n    speaker,\n  }: { speechModel?: DeepgramVoiceConfig; listeningModel?: DeepgramVoiceConfig; speaker?: DeepgramVoiceId } = {}) {\n    const defaultApiKey = process.env.DEEPGRAM_API_KEY;\n\n    const defaultSpeechModel: { name: DeepgramModel; apiKey?: string } = {\n      name: 'aura',\n      apiKey: defaultApiKey,\n    };\n\n    const defaultListeningModel: { name: DeepgramModel; apiKey?: string } = {\n      name: 'nova',\n      apiKey: defaultApiKey,\n    };\n\n    super({\n      speechModel: {\n        name: speechModel?.name ?? defaultSpeechModel.name,\n        apiKey: speechModel?.apiKey ?? defaultSpeechModel.apiKey,\n      },\n      listeningModel: {\n        name: listeningModel?.name ?? defaultListeningModel.name,\n        apiKey: listeningModel?.apiKey ?? defaultListeningModel.apiKey,\n      },\n      speaker,\n    });\n\n    this.storedSpeechModel = {\n      name: speechModel?.name ?? defaultSpeechModel.name,\n      apiKey: speechModel?.apiKey ?? defaultSpeechModel.apiKey,\n    };\n    this.storedListeningModel = {\n      name: listeningModel?.name ?? defaultListeningModel.name,\n      apiKey: listeningModel?.apiKey ?? defaultListeningModel.apiKey,\n    };\n\n    const speechApiKey = speechModel?.apiKey || defaultApiKey;\n    const listeningApiKey = listeningModel?.apiKey || defaultApiKey;\n\n    if (!speechApiKey && !listeningApiKey) {\n      throw new Error('At least one of DEEPGRAM_API_KEY, speechModel.apiKey, or listeningModel.apiKey must be set');\n    }\n\n    if (speechApiKey) {\n      this.speechClient = createClient(speechApiKey);\n    }\n    if (listeningApiKey) {\n      this.listeningClient = createClient(listeningApiKey);\n    }\n\n    this.storedSpeaker = speaker || 'asteria-en';\n  }\n\n  async getSpeakers() {\n    return DEEPGRAM_VOICES.map(voice => ({\n      voiceId: voice,\n    }));\n  }\n\n  async speak(\n    input: string | NodeJS.ReadableStream,\n    options?: {\n      speaker?: string;\n      [key: string]: any;\n    },\n  ): Promise<NodeJS.ReadableStream> {\n    if (!this.speechClient) {\n      throw new Error('Deepgram speech client not configured');\n    }\n\n    let text: string;\n    if (typeof input !== 'string') {\n      const chunks: Buffer[] = [];\n      for await (const chunk of input) {\n        if (typeof chunk === 'string') {\n          chunks.push(Buffer.from(chunk));\n        } else {\n          chunks.push(chunk);\n        }\n      }\n      text = Buffer.concat(chunks).toString('utf-8');\n    } else {\n      text = input;\n    }\n\n    if (text.trim().length === 0) {\n      throw new Error('Input text is empty');\n    }\n\n    const baseModel = this.storedSpeechModel?.name;\n    const speakerId = options?.speaker || this.storedSpeaker;\n\n    const modelName =\n      baseModel && speakerId\n        ? speakerId.startsWith(`${baseModel}-`)\n          ? speakerId\n          : `${baseModel}-${speakerId}`\n        : baseModel || speakerId;\n\n    const speakClient = this.speechClient.speak;\n    const response = await speakClient.request(\n      { text },\n      {\n        model: modelName,\n        ...Object.fromEntries(Object.entries(options ?? {}).filter(([k]) => k !== 'speaker')),\n      },\n    );\n\n    const webStream = await response.getStream();\n    if (!webStream) {\n      throw new Error('No stream returned from Deepgram');\n    }\n\n    const reader = webStream.getReader();\n    const nodeStream = new PassThrough();\n\n    // Add error handling for the stream processing\n    (async () => {\n      try {\n        while (true) {\n          const { done, value } = await reader.read();\n          if (done) {\n            nodeStream.end();\n            break;\n          }\n          nodeStream.write(value);\n        }\n      } catch (error) {\n        nodeStream.destroy(error as Error);\n      }\n    })().catch(error => {\n      nodeStream.destroy(error as Error);\n    });\n\n    return nodeStream;\n  }\n\n  /**\n   * Checks if listening capabilities are enabled.\n   *\n   * @returns {Promise<{ enabled: boolean }>}\n   */\n  async getListener() {\n    return { enabled: true };\n  }\n\n  /**\n   * Transcribes audio with optional speaker diarization.\n   *\n   * @param audioStream - Audio input stream\n   * @param options - Transcription options (diarize, language, etc.)\n   * @returns Promise resolving to:\n   *   - transcript: Full transcript string\n   *   - words: Array of word objects with timing and confidence\n   *   - raw: Complete Deepgram API response\n   *   - speakerSegments: (when diarize=true) Array of {word, speaker, start, end}\n   */\n  async listen(\n    audioStream: NodeJS.ReadableStream,\n    options?: {\n      diarize?: boolean;\n      [key: string]: any;\n    },\n  ): Promise<any> {\n    if (!this.listeningClient) {\n      throw new Error('Deepgram listening client not configured');\n    }\n\n    const chunks: Buffer[] = [];\n    for await (const chunk of audioStream) {\n      if (typeof chunk === 'string') {\n        chunks.push(Buffer.from(chunk));\n      } else {\n        chunks.push(chunk);\n      }\n    }\n    const buffer = Buffer.concat(chunks);\n    const { diarize, diarize_speaker_count: _, ...restOptions } = options ?? {};\n    const { result, error } = await this.listeningClient.listen.prerecorded.transcribeFile(buffer, {\n      ...restOptions,\n      model: this.storedListeningModel?.name,\n      diarize,\n    });\n    if (error) {\n      throw error;\n    }\n\n    const channel = result.results?.channels?.[0];\n    const alt:\n      | {\n          transcript?: string;\n          words?: DeepgramWord[];\n        }\n      | undefined = channel?.alternatives?.[0];\n\n    if (!alt) {\n      return {\n        transcript: '',\n        words: [],\n        raw: result,\n      };\n    }\n\n    const response: any = {\n      transcript: alt.transcript,\n      words: alt.words,\n      raw: result,\n    };\n\n    if (diarize && alt.words) {\n      response.speakerSegments = alt.words.map((w: DeepgramWord) => ({\n        word: w.word,\n        speaker: w.speaker,\n        start: w.start,\n        end: w.end,\n      }));\n    }\n\n    return response;\n  }\n}\n\nexport type { DeepgramVoiceConfig, DeepgramVoiceId, DeepgramModel };\n"]}

package/dist/index.d.ts

@@ -1,4 +1,52 @@
-export { DeepgramVoice } from './_tsup-dts-rollup.js';
-export { DeepgramVoiceConfig } from './_tsup-dts-rollup.js';
-export { DeepgramVoiceId } from './_tsup-dts-rollup.js';
-export { DeepgramModel } from './_tsup-dts-rollup.js';
+import { MastraVoice } from '@mastra/core/voice';
+import type { DeepgramVoiceId, DeepgramModel } from './voices.js';
+interface DeepgramVoiceConfig {
+    name?: DeepgramModel;
+    apiKey?: string;
+    properties?: Record<string, any>;
+    language?: string;
+}
+export declare class DeepgramVoice extends MastraVoice {
+    private speechClient?;
+    private listeningClient?;
+    private storedSpeechModel?;
+    private storedListeningModel?;
+    private storedSpeaker?;
+    constructor({ speechModel, listeningModel, speaker, }?: {
+        speechModel?: DeepgramVoiceConfig;
+        listeningModel?: DeepgramVoiceConfig;
+        speaker?: DeepgramVoiceId;
+    });
+    getSpeakers(): Promise<{
+        voiceId: "asteria-en" | "luna-en" | "stella-en" | "athena-en" | "hera-en" | "orion-en" | "arcas-en" | "perseus-en" | "angus-en" | "orpheus-en" | "helios-en" | "zeus-en";
+    }[]>;
+    speak(input: string | NodeJS.ReadableStream, options?: {
+        speaker?: string;
+        [key: string]: any;
+    }): Promise<NodeJS.ReadableStream>;
+    /**
+     * Checks if listening capabilities are enabled.
+     *
+     * @returns {Promise<{ enabled: boolean }>}
+     */
+    getListener(): Promise<{
+        enabled: boolean;
+    }>;
+    /**
+     * Transcribes audio with optional speaker diarization.
+     *
+     * @param audioStream - Audio input stream
+     * @param options - Transcription options (diarize, language, etc.)
+     * @returns Promise resolving to:
+     *   - transcript: Full transcript string
+     *   - words: Array of word objects with timing and confidence
+     *   - raw: Complete Deepgram API response
+     *   - speakerSegments: (when diarize=true) Array of {word, speaker, start, end}
+     */
+    listen(audioStream: NodeJS.ReadableStream, options?: {
+        diarize?: boolean;
+        [key: string]: any;
+    }): Promise<any>;
+}
+export type { DeepgramVoiceConfig, DeepgramVoiceId, DeepgramModel };
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,WAAW,EAAE,MAAM,oBAAoB,CAAC;AAGjD,OAAO,KAAK,EAAE,eAAe,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AAE/D,UAAU,mBAAmB;IAC3B,IAAI,CAAC,EAAE,aAAa,CAAC;IACrB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IACjC,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAUD,qBAAa,aAAc,SAAQ,WAAW;IAC5C,OAAO,CAAC,YAAY,CAAC,CAAkC;IACvD,OAAO,CAAC,eAAe,CAAC,CAAkC;IAC1D,OAAO,CAAC,iBAAiB,CAAC,CAA2C;IACrE,OAAO,CAAC,oBAAoB,CAAC,CAA2C;IACxE,OAAO,CAAC,aAAa,CAAC,CAAkB;gBAE5B,EACV,WAAW,EACX,cAAc,EACd,OAAO,GACR,GAAE;QAAE,WAAW,CAAC,EAAE,mBAAmB,CAAC;QAAC,cAAc,CAAC,EAAE,mBAAmB,CAAC;QAAC,OAAO,CAAC,EAAE,eAAe,CAAA;KAAO;IAmDxG,WAAW;;;IAMX,KAAK,CACT,KAAK,EAAE,MAAM,GAAG,MAAM,CAAC,cAAc,EACrC,OAAO,CAAC,EAAE;QACR,OAAO,CAAC,EAAE,MAAM,CAAC;QACjB,CAAC,GAAG,EAAE,MAAM,GAAG,GAAG,CAAC;KACpB,GACA,OAAO,CAAC,MAAM,CAAC,cAAc,CAAC;IAwEjC;;;;OAIG;IACG,WAAW;;;IAIjB;;;;;;;;;;OAUG;IACG,MAAM,CACV,WAAW,EAAE,MAAM,CAAC,cAAc,EAClC,OAAO,CAAC,EAAE;QACR,OAAO,CAAC,EAAE,OAAO,CAAC;QAClB,CAAC,GAAG,EAAE,MAAM,GAAG,GAAG,CAAC;KACpB,GACA,OAAO,CAAC,GAAG,CAAC;CAyDhB;AAED,YAAY,EAAE,mBAAmB,EAAE,eAAe,EAAE,aAAa,EAAE,CAAC"}

package/dist/index.js

@@ -24,6 +24,9 @@ var DEEPGRAM_VOICES = [
 var DeepgramVoice = class extends MastraVoice {
   speechClient;
   listeningClient;
+  storedSpeechModel;
+  storedListeningModel;
+  storedSpeaker;
   constructor({
     speechModel,
     listeningModel,
@@ -49,6 +52,14 @@ var DeepgramVoice = class extends MastraVoice {
       },
       speaker
     });
+    this.storedSpeechModel = {
+      name: speechModel?.name ?? defaultSpeechModel.name,
+      apiKey: speechModel?.apiKey ?? defaultSpeechModel.apiKey
+    };
+    this.storedListeningModel = {
+      name: listeningModel?.name ?? defaultListeningModel.name,
+      apiKey: listeningModel?.apiKey ?? defaultListeningModel.apiKey
+    };
     const speechApiKey = speechModel?.apiKey || defaultApiKey;
     const listeningApiKey = listeningModel?.apiKey || defaultApiKey;
     if (!speechApiKey && !listeningApiKey) {
@@ -60,14 +71,12 @@ var DeepgramVoice = class extends MastraVoice {
     if (listeningApiKey) {
       this.listeningClient = createClient(listeningApiKey);
     }
-    this.speaker = speaker || "asteria-en";
+    this.storedSpeaker = speaker || "asteria-en";
   }
   async getSpeakers() {
-    return this.traced(async () => {
-      return DEEPGRAM_VOICES.map((voice) => ({
-        voiceId: voice
-      }));
-    }, "voice.deepgram.getSpeakers")();
+    return DEEPGRAM_VOICES.map((voice) => ({
+      voiceId: voice
+    }));
   }
   async speak(input, options) {
     if (!this.speechClient) {
@@ -90,49 +99,60 @@ var DeepgramVoice = class extends MastraVoice {
     if (text.trim().length === 0) {
       throw new Error("Input text is empty");
     }
-    return this.traced(async () => {
-      if (!this.speechClient) {
-        throw new Error("No speech client configured");
-      }
-      let model;
-      if (options?.speaker) {
-        model = this.speechModel?.name + "-" + options.speaker;
-      } else if (this.speaker) {
-        model = this.speechModel?.name + "-" + this.speaker;
-      }
-      const speakClient = this.speechClient.speak;
-      const response = await speakClient.request(
-        { text },
-        {
-          model,
-          ...options
-        }
-      );
-      const webStream = await response.getStream();
-      if (!webStream) {
-        throw new Error("No stream returned from Deepgram");
+    const baseModel = this.storedSpeechModel?.name;
+    const speakerId = options?.speaker || this.storedSpeaker;
+    const modelName = baseModel && speakerId ? speakerId.startsWith(`${baseModel}-`) ? speakerId : `${baseModel}-${speakerId}` : baseModel || speakerId;
+    const speakClient = this.speechClient.speak;
+    const response = await speakClient.request(
+      { text },
+      {
+        model: modelName,
+        ...Object.fromEntries(Object.entries(options ?? {}).filter(([k]) => k !== "speaker"))
       }
-      const reader = webStream.getReader();
-      const nodeStream = new PassThrough();
-      (async () => {
-        try {
-          while (true) {
-            const { done, value } = await reader.read();
-            if (done) {
-              nodeStream.end();
-              break;
-            }
-            nodeStream.write(value);
+    );
+    const webStream = await response.getStream();
+    if (!webStream) {
+      throw new Error("No stream returned from Deepgram");
+    }
+    const reader = webStream.getReader();
+    const nodeStream = new PassThrough();
+    (async () => {
+      try {
+        while (true) {
+          const { done, value } = await reader.read();
+          if (done) {
+            nodeStream.end();
+            break;
           }
-        } catch (error) {
-          nodeStream.destroy(error);
+          nodeStream.write(value);
         }
-      })().catch((error) => {
+      } catch (error) {
         nodeStream.destroy(error);
-      });
-      return nodeStream;
-    }, "voice.deepgram.speak")();
+      }
+    })().catch((error) => {
+      nodeStream.destroy(error);
+    });
+    return nodeStream;
+  }
+  /**
+   * Checks if listening capabilities are enabled.
+   *
+   * @returns {Promise<{ enabled: boolean }>}
+   */
+  async getListener() {
+    return { enabled: true };
   }
+  /**
+   * Transcribes audio with optional speaker diarization.
+   *
+   * @param audioStream - Audio input stream
+   * @param options - Transcription options (diarize, language, etc.)
+   * @returns Promise resolving to:
+   *   - transcript: Full transcript string
+   *   - words: Array of word objects with timing and confidence
+   *   - raw: Complete Deepgram API response
+   *   - speakerSegments: (when diarize=true) Array of {word, speaker, start, end}
+   */
   async listen(audioStream, options) {
     if (!this.listeningClient) {
       throw new Error("Deepgram listening client not configured");
@@ -146,24 +166,41 @@ var DeepgramVoice = class extends MastraVoice {
       }
     }
     const buffer = Buffer.concat(chunks);
-    return this.traced(async () => {
-      if (!this.listeningClient) {
-        throw new Error("No listening client configured");
-      }
-      const { result, error } = await this.listeningClient.listen.prerecorded.transcribeFile(buffer, {
-        model: this.listeningModel?.name,
-        ...options
-      });
-      if (error) {
-        throw error;
-      }
-      const transcript = result.results?.channels?.[0]?.alternatives?.[0]?.transcript;
-      if (!transcript) {
-        throw new Error("No transcript found in Deepgram response");
-      }
-      return transcript;
-    }, "voice.deepgram.listen")();
+    const { diarize, diarize_speaker_count: _, ...restOptions } = options ?? {};
+    const { result, error } = await this.listeningClient.listen.prerecorded.transcribeFile(buffer, {
+      ...restOptions,
+      model: this.storedListeningModel?.name,
+      diarize
+    });
+    if (error) {
+      throw error;
+    }
+    const channel = result.results?.channels?.[0];
+    const alt = channel?.alternatives?.[0];
+    if (!alt) {
+      return {
+        transcript: "",
+        words: [],
+        raw: result
+      };
+    }
+    const response = {
+      transcript: alt.transcript,
+      words: alt.words,
+      raw: result
+    };
+    if (diarize && alt.words) {
+      response.speakerSegments = alt.words.map((w) => ({
+        word: w.word,
+        speaker: w.speaker,
+        start: w.start,
+        end: w.end
+      }));
+    }
+    return response;
   }
 };
 
 export { DeepgramVoice };
+//# sourceMappingURL=index.js.map
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/voices.ts","../src/index.ts"],"names":[],"mappings":";;;;;;;AAKO,IAAM,eAAA,GAAkB;AAAA,EAC7B,YAAA;AAAA,EACA,SAAA;AAAA,EACA,WAAA;AAAA,EACA,WAAA;AAAA,EACA,SAAA;AAAA,EACA,UAAA;AAAA,EACA,UAAA;AAAA,EACA,YAAA;AAAA,EACA,UAAA;AAAA,EACA,YAAA;AAAA,EACA,WAAA;AAAA,EACA;AACF,CAAA;;;ACKO,IAAM,aAAA,GAAN,cAA4B,WAAA,CAAY;AAAA,EACrC,YAAA;AAAA,EACA,eAAA;AAAA,EACA,iBAAA;AAAA,EACA,oBAAA;AAAA,EACA,aAAA;AAAA,EAER,WAAA,CAAY;AAAA,IACV,WAAA;AAAA,IACA,cAAA;AAAA,IACA;AAAA,GACF,GAA4G,EAAC,EAAG;AAC9G,IAAA,MAAM,aAAA,GAAgB,QAAQ,GAAA,CAAI,gBAAA;AAElC,IAAA,MAAM,kBAAA,GAA+D;AAAA,MACnE,IAAA,EAAM,MAAA;AAAA,MACN,MAAA,EAAQ;AAAA,KACV;AAEA,IAAA,MAAM,qBAAA,GAAkE;AAAA,MACtE,IAAA,EAAM,MAAA;AAAA,MACN,MAAA,EAAQ;AAAA,KACV;AAEA,IAAA,KAAA,CAAM;AAAA,MACJ,WAAA,EAAa;AAAA,QACX,IAAA,EAAM,WAAA,EAAa,IAAA,IAAQ,kBAAA,CAAmB,IAAA;AAAA,QAC9C,MAAA,EAAQ,WAAA,EAAa,MAAA,IAAU,kBAAA,CAAmB;AAAA,OACpD;AAAA,MACA,cAAA,EAAgB;AAAA,QACd,IAAA,EAAM,cAAA,EAAgB,IAAA,IAAQ,qBAAA,CAAsB,IAAA;AAAA,QACpD,MAAA,EAAQ,cAAA,EAAgB,MAAA,IAAU,qBAAA,CAAsB;AAAA,OAC1D;AAAA,MACA;AAAA,KACD,CAAA;AAED,IAAA,IAAA,CAAK,iBAAA,GAAoB;AAAA,MACvB,IAAA,EAAM,WAAA,EAAa,IAAA,IAAQ,kBAAA,CAAmB,IAAA;AAAA,MAC9C,MAAA,EAAQ,WAAA,EAAa,MAAA,IAAU,kBAAA,CAAmB;AAAA,KACpD;AACA,IAAA,IAAA,CAAK,oBAAA,GAAuB;AAAA,MAC1B,IAAA,EAAM,cAAA,EAAgB,IAAA,IAAQ,qBAAA,CAAsB,IAAA;AAAA,MACpD,MAAA,EAAQ,cAAA,EAAgB,MAAA,IAAU,qBAAA,CAAsB;AAAA,KAC1D;AAEA,IAAA,MAAM,YAAA,GAAe,aAAa,MAAA,IAAU,aAAA;AAC5C,IAAA,MAAM,eAAA,GAAkB,gBAAgB,MAAA,IAAU,aAAA;AAElD,IAAA,IAAI,CAAC,YAAA,IAAgB,CAAC,eAAA,EAAiB;AACrC,MAAA,MAAM,IAAI,MAAM,4FAA4F,CAAA;AAAA,IAC9G;AAEA,IAAA,IAAI,YAAA,EAAc;AAChB,MAAA,IAAA,CAAK,YAAA,GAAe,aAAa,YAAY,CAAA;AAAA,IAC/C;AACA,IAAA,IAAI,eAAA,EAAiB;AACnB,MAAA,IAAA,CAAK,eAAA,GAAkB,aAAa,eAAe,CAAA;AAAA,IACrD;AAEA,IAAA,IAAA,CAAK,gBAAgB,OAAA,IAAW,YAAA;AAAA,EAClC;AAAA,EAEA,MAAM,WAAA,GAAc;AAClB,IAAA,OAAO,eAAA,CAAgB,IAAI,CAAA,KAAA,MAAU;AAAA,MACnC,OAAA,EAAS;AAAA,KACX,CAAE,CAAA;AAAA,EACJ;AAAA,EAEA,MAAM,KAAA,CACJ,KAAA,EACA,OAAA,EAIgC;AAChC,IAAA,IAAI,CAAC,KAAK,YAAA,EAAc;AACtB,MAAA,MAAM,IAAI,MAAM,uCAAuC,CAAA;AAAA,IACzD;AAEA,IAAA,IAAI,IAAA;AACJ,IAAA,IAAI,OAAO,UAAU,QAAA,EAAU;AAC7B,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,WAAA,MAAiB,SAAS,KAAA,EAAO;AAC/B,QAAA,IAAI,OAAO,UAAU,QAAA,EAAU;AAC7B,UAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,KAAK,CAAC,CAAA;AAAA,QAChC,CAAA,MAAO;AACL,UAAA,MAAA,CAAO,KAAK,KAAK,CAAA;AAAA,QACnB;AAAA,MACF;AACA,MAAA,IAAA,GAAO,MAAA,CAAO,MAAA,CAAO,MAAM,CAAA,CAAE,SAAS,OAAO,CAAA;AAAA,IAC/C,CAAA,MAAO;AACL,MAAA,IAAA,GAAO,KAAA;AAAA,IACT;AAEA,IAAA,IAAI,IAAA,CAAK,IAAA,EAAK,CAAE,MAAA,KAAW,CAAA,EAAG;AAC5B,MAAA,MAAM,IAAI,MAAM,qBAAqB,CAAA;AAAA,IACvC;AAEA,IAAA,MAAM,SAAA,GAAY,KAAK,iBAAA,EAAmB,IAAA;AAC1C,IAAA,MAAM,SAAA,GAAY,OAAA,EAAS,OAAA,IAAW,IAAA,CAAK,aAAA;AAE3C,IAAA,MAAM,SAAA,GACJ,SAAA,IAAa,SAAA,GACT,SAAA,CAAU,WAAW,CAAA,EAAG,SAAS,CAAA,CAAA,CAAG,CAAA,GAClC,YACA,CAAA,EAAG,SAAS,CAAA,CAAA,EAAI,SAAS,KAC3B,SAAA,IAAa,SAAA;AAEnB,IAAA,MAAM,WAAA,GAAc,KAAK,YAAA,CAAa,KAAA;AACtC,IAAA,MAAM,QAAA,GAAW,MAAM,WAAA,CAAY,OAAA;AAAA,MACjC,EAAE,IAAA,EAAK;AAAA,MACP;AAAA,QACE,KAAA,EAAO,SAAA;AAAA,QACP,GAAG,MAAA,CAAO,WAAA,CAAY,MAAA,CAAO,OAAA,CAAQ,WAAW,EAAE,CAAA,CAAE,MAAA,CAAO,CAAC,CAAC,CAAC,CAAA,KAAM,CAAA,KAAM,SAAS,CAAC;AAAA;AACtF,KACF;AAEA,IAAA,MAAM,SAAA,GAAY,MAAM,QAAA,CAAS,SAAA,EAAU;AAC3C,IAAA,IAAI,CAAC,SAAA,EAAW;AACd,MAAA,MAAM,IAAI,MAAM,kCAAkC,CAAA;AAAA,IACpD;AAEA,IAAA,MAAM,MAAA,GAAS,UAAU,SAAA,EAAU;AACnC,IAAA,MAAM,UAAA,GAAa,IAAI,WAAA,EAAY;AAGnC,IAAA,CAAC,YAAY;AACX,MAAA,IAAI;AACF,QAAA,OAAO,IAAA,EAAM;AACX,UAAA,MAAM,EAAE,IAAA,EAAM,KAAA,EAAM,GAAI,MAAM,OAAO,IAAA,EAAK;AAC1C,UAAA,IAAI,IAAA,EAAM;AACR,YAAA,UAAA,CAAW,GAAA,EAAI;AACf,YAAA;AAAA,UACF;AACA,UAAA,UAAA,CAAW,MAAM,KAAK,CAAA;AAAA,QACxB;AAAA,MACF,SAAS,KAAA,EAAO;AACd,QAAA,UAAA,CAAW,QAAQ,KAAc,CAAA;AAAA,MACnC;AAAA,IACF,CAAA,GAAG,CAAE,KAAA,CAAM,CAAA,KAAA,KAAS;AAClB,MAAA,UAAA,CAAW,QAAQ,KAAc,CAAA;AAAA,IACnC,CAAC,CAAA;AAED,IAAA,OAAO,UAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,MAAM,WAAA,GAAc;AAClB,IAAA,OAAO,EAAE,SAAS,IAAA,EAAK;AAAA,EACzB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,MAAM,MAAA,CACJ,WAAA,EACA,OAAA,EAIc;AACd,IAAA,IAAI,CAAC,KAAK,eAAA,EAAiB;AACzB,MAAA,MAAM,IAAI,MAAM,0CAA0C,CAAA;AAAA,IAC5D;AAEA,IAAA,MAAM,SAAmB,EAAC;AAC1B,IAAA,WAAA,MAAiB,SAAS,WAAA,EAAa;AACrC,MAAA,IAAI,OAAO,UAAU,QAAA,EAAU;AAC7B,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,KAAK,CAAC,CAAA;AAAA,MAChC,CAAA,MAAO;AACL,QAAA,MAAA,CAAO,KAAK,KAAK,CAAA;AAAA,MACnB;AAAA,IACF;AACA,IAAA,MAAM,MAAA,GAAS,MAAA,CAAO,MAAA,CAAO,MAAM,CAAA;AACnC,IAAA,MAAM,EAAE,SAAS,qBAAA,EAAuB,CAAA,EAAG,GAAG,WAAA,EAAY,GAAI,WAAW,EAAC;AAC1E,IAAA,MAAM,EAAE,MAAA,EAAQ,KAAA,EAAM,GAAI,MAAM,KAAK,eAAA,CAAgB,MAAA,CAAO,WAAA,CAAY,cAAA,CAAe,MAAA,EAAQ;AAAA,MAC7F,GAAG,WAAA;AAAA,MACH,KAAA,EAAO,KAAK,oBAAA,EAAsB,IAAA;AAAA,MAClC;AAAA,KACD,CAAA;AACD,IAAA,IAAI,KAAA,EAAO;AACT,MAAA,MAAM,KAAA;AAAA,IACR;AAEA,IAAA,MAAM,OAAA,GAAU,MAAA,CAAO,OAAA,EAAS,QAAA,GAAW,CAAC,CAAA;AAC5C,IAAA,MAAM,GAAA,GAKU,OAAA,EAAS,YAAA,GAAe,CAAC,CAAA;AAEzC,IAAA,IAAI,CAAC,GAAA,EAAK;AACR,MAAA,OAAO;AAAA,QACL,UAAA,EAAY,EAAA;AAAA,QACZ,OAAO,EAAC;AAAA,QACR,GAAA,EAAK;AAAA,OACP;AAAA,IACF;AAEA,IAAA,MAAM,QAAA,GAAgB;AAAA,MACpB,YAAY,GAAA,CAAI,UAAA;AAAA,MAChB,OAAO,GAAA,CAAI,KAAA;AAAA,MACX,GAAA,EAAK;AAAA,KACP;AAEA,IAAA,IAAI,OAAA,IAAW,IAAI,KAAA,EAAO;AACxB,MAAA,QAAA,CAAS,eAAA,GAAkB,GAAA,CAAI,KAAA,CAAM,GAAA,CAAI,CAAC,CAAA,MAAqB;AAAA,QAC7D,MAAM,CAAA,CAAE,IAAA;AAAA,QACR,SAAS,CAAA,CAAE,OAAA;AAAA,QACX,OAAO,CAAA,CAAE,KAAA;AAAA,QACT,KAAK,CAAA,CAAE;AAAA,OACT,CAAE,CAAA;AAAA,IACJ;AAEA,IAAA,OAAO,QAAA;AAAA,EACT;AACF","file":"index.js","sourcesContent":["/**\n * List of available Deepgram voice models for text-to-speech\n * Each voice is designed for specific use cases and languages\n * Format: {name}-{language} (e.g. asteria-en)\n */\nexport const DEEPGRAM_VOICES = [\n  'asteria-en',\n  'luna-en',\n  'stella-en',\n  'athena-en',\n  'hera-en',\n  'orion-en',\n  'arcas-en',\n  'perseus-en',\n  'angus-en',\n  'orpheus-en',\n  'helios-en',\n  'zeus-en',\n] as const;\n\nexport type DeepgramVoiceId = (typeof DEEPGRAM_VOICES)[number];\n\n/**\n * List of available Deepgram models for text-to-speech and speech-to-text\n */\nexport const DEEPGRAM_MODELS = ['aura', 'whisper', 'base', 'enhanced', 'nova', 'nova-2', 'nova-3'] as const;\n\nexport type DeepgramModel = (typeof DEEPGRAM_MODELS)[number];\n","import { PassThrough } from 'node:stream';\n\nimport { createClient } from '@deepgram/sdk';\nimport { MastraVoice } from '@mastra/core/voice';\n\nimport { DEEPGRAM_VOICES } from './voices';\nimport type { DeepgramVoiceId, DeepgramModel } from './voices';\n\ninterface DeepgramVoiceConfig {\n  name?: DeepgramModel;\n  apiKey?: string;\n  properties?: Record<string, any>;\n  language?: string;\n}\n\ninterface DeepgramWord {\n  word: string;\n  start?: number;\n  end?: number;\n  confidence?: number;\n  speaker?: number;\n}\n\nexport class DeepgramVoice extends MastraVoice {\n  private speechClient?: ReturnType<typeof createClient>;\n  private listeningClient?: ReturnType<typeof createClient>;\n  private storedSpeechModel?: { name: DeepgramModel; apiKey?: string };\n  private storedListeningModel?: { name: DeepgramModel; apiKey?: string };\n  private storedSpeaker?: DeepgramVoiceId;\n\n  constructor({\n    speechModel,\n    listeningModel,\n    speaker,\n  }: { speechModel?: DeepgramVoiceConfig; listeningModel?: DeepgramVoiceConfig; speaker?: DeepgramVoiceId } = {}) {\n    const defaultApiKey = process.env.DEEPGRAM_API_KEY;\n\n    const defaultSpeechModel: { name: DeepgramModel; apiKey?: string } = {\n      name: 'aura',\n      apiKey: defaultApiKey,\n    };\n\n    const defaultListeningModel: { name: DeepgramModel; apiKey?: string } = {\n      name: 'nova',\n      apiKey: defaultApiKey,\n    };\n\n    super({\n      speechModel: {\n        name: speechModel?.name ?? defaultSpeechModel.name,\n        apiKey: speechModel?.apiKey ?? defaultSpeechModel.apiKey,\n      },\n      listeningModel: {\n        name: listeningModel?.name ?? defaultListeningModel.name,\n        apiKey: listeningModel?.apiKey ?? defaultListeningModel.apiKey,\n      },\n      speaker,\n    });\n\n    this.storedSpeechModel = {\n      name: speechModel?.name ?? defaultSpeechModel.name,\n      apiKey: speechModel?.apiKey ?? defaultSpeechModel.apiKey,\n    };\n    this.storedListeningModel = {\n      name: listeningModel?.name ?? defaultListeningModel.name,\n      apiKey: listeningModel?.apiKey ?? defaultListeningModel.apiKey,\n    };\n\n    const speechApiKey = speechModel?.apiKey || defaultApiKey;\n    const listeningApiKey = listeningModel?.apiKey || defaultApiKey;\n\n    if (!speechApiKey && !listeningApiKey) {\n      throw new Error('At least one of DEEPGRAM_API_KEY, speechModel.apiKey, or listeningModel.apiKey must be set');\n    }\n\n    if (speechApiKey) {\n      this.speechClient = createClient(speechApiKey);\n    }\n    if (listeningApiKey) {\n      this.listeningClient = createClient(listeningApiKey);\n    }\n\n    this.storedSpeaker = speaker || 'asteria-en';\n  }\n\n  async getSpeakers() {\n    return DEEPGRAM_VOICES.map(voice => ({\n      voiceId: voice,\n    }));\n  }\n\n  async speak(\n    input: string | NodeJS.ReadableStream,\n    options?: {\n      speaker?: string;\n      [key: string]: any;\n    },\n  ): Promise<NodeJS.ReadableStream> {\n    if (!this.speechClient) {\n      throw new Error('Deepgram speech client not configured');\n    }\n\n    let text: string;\n    if (typeof input !== 'string') {\n      const chunks: Buffer[] = [];\n      for await (const chunk of input) {\n        if (typeof chunk === 'string') {\n          chunks.push(Buffer.from(chunk));\n        } else {\n          chunks.push(chunk);\n        }\n      }\n      text = Buffer.concat(chunks).toString('utf-8');\n    } else {\n      text = input;\n    }\n\n    if (text.trim().length === 0) {\n      throw new Error('Input text is empty');\n    }\n\n    const baseModel = this.storedSpeechModel?.name;\n    const speakerId = options?.speaker || this.storedSpeaker;\n\n    const modelName =\n      baseModel && speakerId\n        ? speakerId.startsWith(`${baseModel}-`)\n          ? speakerId\n          : `${baseModel}-${speakerId}`\n        : baseModel || speakerId;\n\n    const speakClient = this.speechClient.speak;\n    const response = await speakClient.request(\n      { text },\n      {\n        model: modelName,\n        ...Object.fromEntries(Object.entries(options ?? {}).filter(([k]) => k !== 'speaker')),\n      },\n    );\n\n    const webStream = await response.getStream();\n    if (!webStream) {\n      throw new Error('No stream returned from Deepgram');\n    }\n\n    const reader = webStream.getReader();\n    const nodeStream = new PassThrough();\n\n    // Add error handling for the stream processing\n    (async () => {\n      try {\n        while (true) {\n          const { done, value } = await reader.read();\n          if (done) {\n            nodeStream.end();\n            break;\n          }\n          nodeStream.write(value);\n        }\n      } catch (error) {\n        nodeStream.destroy(error as Error);\n      }\n    })().catch(error => {\n      nodeStream.destroy(error as Error);\n    });\n\n    return nodeStream;\n  }\n\n  /**\n   * Checks if listening capabilities are enabled.\n   *\n   * @returns {Promise<{ enabled: boolean }>}\n   */\n  async getListener() {\n    return { enabled: true };\n  }\n\n  /**\n   * Transcribes audio with optional speaker diarization.\n   *\n   * @param audioStream - Audio input stream\n   * @param options - Transcription options (diarize, language, etc.)\n   * @returns Promise resolving to:\n   *   - transcript: Full transcript string\n   *   - words: Array of word objects with timing and confidence\n   *   - raw: Complete Deepgram API response\n   *   - speakerSegments: (when diarize=true) Array of {word, speaker, start, end}\n   */\n  async listen(\n    audioStream: NodeJS.ReadableStream,\n    options?: {\n      diarize?: boolean;\n      [key: string]: any;\n    },\n  ): Promise<any> {\n    if (!this.listeningClient) {\n      throw new Error('Deepgram listening client not configured');\n    }\n\n    const chunks: Buffer[] = [];\n    for await (const chunk of audioStream) {\n      if (typeof chunk === 'string') {\n        chunks.push(Buffer.from(chunk));\n      } else {\n        chunks.push(chunk);\n      }\n    }\n    const buffer = Buffer.concat(chunks);\n    const { diarize, diarize_speaker_count: _, ...restOptions } = options ?? {};\n    const { result, error } = await this.listeningClient.listen.prerecorded.transcribeFile(buffer, {\n      ...restOptions,\n      model: this.storedListeningModel?.name,\n      diarize,\n    });\n    if (error) {\n      throw error;\n    }\n\n    const channel = result.results?.channels?.[0];\n    const alt:\n      | {\n          transcript?: string;\n          words?: DeepgramWord[];\n        }\n      | undefined = channel?.alternatives?.[0];\n\n    if (!alt) {\n      return {\n        transcript: '',\n        words: [],\n        raw: result,\n      };\n    }\n\n    const response: any = {\n      transcript: alt.transcript,\n      words: alt.words,\n      raw: result,\n    };\n\n    if (diarize && alt.words) {\n      response.speakerSegments = alt.words.map((w: DeepgramWord) => ({\n        word: w.word,\n        speaker: w.speaker,\n        start: w.start,\n        end: w.end,\n      }));\n    }\n\n    return response;\n  }\n}\n\nexport type { DeepgramVoiceConfig, DeepgramVoiceId, DeepgramModel };\n"]}

package/dist/voices.d.ts

@@ -0,0 +1,13 @@
+/**
+ * List of available Deepgram voice models for text-to-speech
+ * Each voice is designed for specific use cases and languages
+ * Format: {name}-{language} (e.g. asteria-en)
+ */
+export declare const DEEPGRAM_VOICES: readonly ["asteria-en", "luna-en", "stella-en", "athena-en", "hera-en", "orion-en", "arcas-en", "perseus-en", "angus-en", "orpheus-en", "helios-en", "zeus-en"];
+export type DeepgramVoiceId = (typeof DEEPGRAM_VOICES)[number];
+/**
+ * List of available Deepgram models for text-to-speech and speech-to-text
+ */
+export declare const DEEPGRAM_MODELS: readonly ["aura", "whisper", "base", "enhanced", "nova", "nova-2", "nova-3"];
+export type DeepgramModel = (typeof DEEPGRAM_MODELS)[number];
+//# sourceMappingURL=voices.d.ts.map

package/dist/voices.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"voices.d.ts","sourceRoot":"","sources":["../src/voices.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH,eAAO,MAAM,eAAe,iKAalB,CAAC;AAEX,MAAM,MAAM,eAAe,GAAG,CAAC,OAAO,eAAe,CAAC,CAAC,MAAM,CAAC,CAAC;AAE/D;;GAEG;AACH,eAAO,MAAM,eAAe,8EAA+E,CAAC;AAE5G,MAAM,MAAM,aAAa,GAAG,CAAC,OAAO,eAAe,CAAC,CAAC,MAAM,CAAC,CAAC"}