Speech MCP Server

Instructions

Implementation Reference

• src/index.ts:294-307 (handler)

  Handler logic for executing the text_to_speech_with_options tool. Extracts arguments and delegates to TTSClient.generateAndPlayAudio method.
  case "text_to_speech_with_options": { const args = request.params.arguments as unknown as TextToSpeechWithOptionsArgs; if (!args.text) { throw new Error("Missing required argument: text"); } await ttsClient.generateAndPlayAudio(args.text, args.voice, args.speed); return { content: [{ type: "text", text: `Successfully generated and played audio${args.voice ? ` using voice: ${args.voice}` : ''} (speed: ${args.speed || 1.0})` }], }; }
• src/index.ts:61-86 (schema)

  Tool definition including name, description, and input schema for validation.
  const textToSpeechWithOptionsTool: Tool = { name: "text_to_speech_with_options", description: "Convert text to speech with customizable speed", inputSchema: { type: "object", properties: { text: { type: "string", description: "The text to convert to speech", minLength: 1, maxLength: 1000, }, voice: { type: "string", description: "The voice to use for speech synthesis (e.g. 'af_bella'). Use list_voices to see available options.", }, speed: { type: "number", description: "Speech rate multiplier (0.5 to 2.0)", minimum: 0.5, maximum: 2.0, }, }, required: ["text"], }, };
• src/index.ts:34-37 (schema)

  TypeScript interface defining the expected arguments for the tool.
  interface TextToSpeechWithOptionsArgs extends TextToSpeechArgs { speed?: number; voice?: KokoroVoice; }
• src/index.ts:356-360 (registration)

  Registers the textToSpeechWithOptionsTool in the list returned by ListToolsRequest handler.
  textToSpeechTool, textToSpeechWithOptionsTool, listVoicesTool, getModelStatusTool, ],
• src/index.ts:230-250 (helper)

  Core helper method in TTSClient that performs the actual text-to-speech generation and playback using KokoroTTS.
  async generateAndPlayAudio(text: string, voice?: KokoroVoice, speed?: number): Promise<void> { await this.waitForInit(); if (!this.ttsInstance) { throw new Error("TTS model not initialized"); } const audio = await this.ttsInstance.generate(text, { voice: voice || DEFAULT_VOICE, // @ts-ignore-line speed: speed || DEFAULT_SPEECH_SPEED, }); const tempFile = join(tmpdir(), `${Date.now()}.wav`); await audio.save(tempFile); await player.play({ path: tempFile, sync: true }); } }