ElevenLabs MCP Server

Instructions

Voice preview files are saved as: voice_design_(generated_voice_id)_(timestamp).mp3

Example file name: voice_design_Ya2J5uIa5Pq14DNPsbC1_20250403_164949.mp3

⚠️ COST WARNING: This tool makes an API call to ElevenLabs which may incur costs. Only use when explicitly requested by the user.

Implementation Reference

• elevenlabs_mcp/server.py:886-920 (handler)

  The handler function that implements the 'text_to_voice' tool. It generates three voice previews from a description using the ElevenLabs API, saves the MP3 audio files, and returns the file paths and generated voice IDs.

  def text_to_voice(
      voice_description: str,
      text: str | None = None,
      output_directory: str | None = None,
  ) -> TextContent:
      if voice_description == "":
          make_error("Voice description is required.")
  
      previews = client.text_to_voice.create_previews(
          voice_description=voice_description,
          text=text,
          auto_generate_text=True if text is None else False,
      )
  
      output_path = make_output_path(output_directory, base_path)
  
      generated_voice_ids = []
      output_file_paths = []
  
      for preview in previews.previews:
          output_file_name = make_output_file(
              "voice_design", preview.generated_voice_id, output_path, "mp3", full_id=True
          )
          output_file_paths.append(str(output_file_name))
          generated_voice_ids.append(preview.generated_voice_id)
          audio_bytes = base64.b64decode(preview.audio_base_64)
  
          with open(output_path / output_file_name, "wb") as f:
              f.write(audio_bytes)
  
      return TextContent(
          type="text",
          text=f"Success. Files saved at: {', '.join(output_file_paths)}. Generated voice IDs are: {', '.join(generated_voice_ids)}",
      )

• elevenlabs_mcp/server.py:876-885 (registration)

  Registers the 'text_to_voice' tool with the MCP server using the @mcp.tool decorator, providing a detailed description that serves as the tool schema.

  @mcp.tool(
      description="""Create voice previews from a text prompt. Creates three previews with slight variations. Saves the previews to a given directory. If no text is provided, the tool will auto-generate text.
  
      Voice preview files are saved as: voice_design_(generated_voice_id)_(timestamp).mp3
  
      Example file name: voice_design_Ya2J5uIa5Pq14DNPsbC1_20250403_164949.mp3
  
      ⚠️ COST WARNING: This tool makes an API call to ElevenLabs which may incur costs. Only use when explicitly requested by the user.
      """
  )

• elevenlabs_mcp/server.py:922-943 (helper)

  Supporting tool 'create_voice_from_preview' that uses the generated voice ID from 'text_to_voice' to add the voice to the library.

  @mcp.tool(
      description="""Add a generated voice to the voice library. Uses the voice ID from the `text_to_voice` tool.
  
      ⚠️ COST WARNING: This tool makes an API call to ElevenLabs which may incur costs. Only use when explicitly requested by the user.
      """
  )
  def create_voice_from_preview(
      generated_voice_id: str,
      voice_name: str,
      voice_description: str,
  ) -> TextContent:
      voice = client.text_to_voice.create_voice_from_preview(
          voice_name=voice_name,
          voice_description=voice_description,
          generated_voice_id=generated_voice_id,
      )
  
      return TextContent(
          type="text",
          text=f"Success. Voice created: {voice.name} with ID:{voice.voice_id}",
      )