Frequently Asked Questions about Speech MCP Server

Q: What is the Speech MCP Server?

A: The Speech MCP Server is a Model Context Protocol (MCP) server that provides text-to-speech capabilities using the Kokoro TTS model. It allows AI agents to convert text into high-quality speech.

Q: What is MCP?

A: MCP stands for Model Context Protocol. It’s an open protocol that standardizes how applications provide context to LLMs, enabling AI models to interact with external data sources and tools.

Q: How do I install the Speech MCP Server?

A: You can install the Speech MCP Server using npm, pnpm, or yarn. The commands are npm install speech-mcp-server, pnpm add speech-mcp-server, or yarn add speech-mcp-server.

Q: What environment variables can I configure?

A: You can configure MCP_DEFAULT_SPEECH_SPEED to set the default speech speed multiplier (range: 0.5 to 2.0) and MCP_DEFAULT_VOICE to set the default voice for text-to-speech.

Q: How do I run the server?

A: After installation, run the server using the command npm start. You can also customize the configuration using environment variables before running the command, such as MCP_DEFAULT_SPEECH_SPEED=1.5 MCP_DEFAULT_VOICE=af_bella npm start.

Q: What tools are available in the Speech MCP Server?

A: The server provides the following MCP tools: text_to_speech, text_to_speech_with_options, list_voices, and get_model_status.

Q: What does the text_to_speech tool do?

A: The text_to_speech tool converts text to speech using the default server settings. You provide the text, and the server generates the speech output.

Q: How do I customize the speech output?

A: Use the text_to_speech_with_options tool. It allows you to specify the voice, speed, and other parameters to customize the speech output.

Q: How do I list all available voices?

A: Use the list_voices tool. It returns a list of all voices supported by the server.

Q: How do I check the status of the TTS model initialization?

A: Use the get_model_status tool. It’s particularly useful when starting the server for the first time to ensure the model is properly initialized.

Q: What possible status values can get_model_status return?

A: Possible status values include: uninitialized, initializing, ready, and error.

Q: How can I integrate the Speech MCP Server with Claude Desktop?

A: Add the following to your Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json):

{ “servers”: { “speech”: { “command”: “npx”, “args”: [“@decodershq/speech-mcp-server”] } } }

Q: What should I do if I encounter model initialization issues?

A: The server automatically retries initialization up to 3 times. Use get_model_status to monitor progress. If it fails, try manually removing model files and restarting the server. The command would be rm -rf ~/.npm/_npx/**/node_modules/@huggingface/transformers/.cache/onnx-community/Kokoro-82M-v1.0-ONNX/onnx/model_quantized.onnx rm -rf ~/.cache/huggingface/transformers/onnx-community/Kokoro-82M-v1.0-ONNX/onnx/model_quantized.onnx.

Q: Where can I find more information about UBOS?

A: UBOS is a full-stack AI Agent Development Platform focused on bringing AI Agents to every business department. Visit https://ubos.tech for more information.