home / mcp / say mcp server

Say MCP Server

MCP server for macOS text-to-speech functionality

Installation
Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "bmorphism-say-mcp-server": {
      "command": "node",
      "args": [
        "/path/to/say-mcp-server/build/index.js"
      ]
    }
  }
}

You can run a macOS-native text-to-speech service inside your MCP environment. This MCP server uses the system’s built‑in say command to speak text aloud, and it exposes practical tools for reading content, notes, transcripts, and search results with configurable voices, rates, and emphasis.

How to use

Use the MCP client to call the say MCP server and request spoken output. You can send text to be spoken, optionally choosing a voice and rate to tailor the narration. You can also run speech in the background so you can continue interacting with other MCP actions while speech plays.

How to install

Prerequisites: You are on macOS and have Node.js installed (Node.js 14 or newer). If you don’t have Node.js, install it from the official source for your macOS version.

How to install (continued)

Install the MCP server package using npm.

How to install (continued)

npm install say-mcp-server

How to configure the MCP server

Add an MCP server entry to your MCP settings so the client can connect to the local stdio server provided by this package.

How to configure the MCP server (continued)

{
  "mcpServers": {
    "say": {
      "command": "node",
      "args": ["/path/to/say-mcp-server/build/index.js"]
    }
  }
}

Notes on usage patterns

- Use the speak tool to read text aloud with optional voice, rate, and background execution. - Use list_voices to discover which macOS voices are available on your system. - Combine speech with other MCP tools to read results from searches or transcripts.

Troubleshooting

If you don’t hear speech, verify that macOS VoiceOver or other accessibility features are not interfering, ensure the say command is accessible, and confirm the Node process is running the provided script. Check that your MCP client is sending text to the say server and that the server is reachable via the configured stdio channel.

Additional configuration and security considerations

Only expose the MCP server to trusted clients in your local network. Do not place it on public endpoints without proper authentication. Keep the system voice configuration consistent across clients to avoid unexpected noise levels.

Tools

This server exposes two practical tools for speaking text and listing available voices.

Available tools

- speak: Speak provided text aloud with optional voice, rate, and background execution. Supports in-text directives for volume, rate, emphasis, and pitch.

- list_voices: List all available text-to-speech voices on the system.

Available tools

speak

Speaks input text aloud with optional voice, rate, background mode, and in-text directives for volume, rate, emphasis, and pitch.

list_voices

Lists all available macOS TTS voices on the system.