macOS Say TTS Server

Example Use Cases

• 1

  Read search results or summaries aloud to users using a native macOS voice.

• 2

  Vocalize transcripts, notes, or long-form text with fine-grained control over rate, volume, pitch, and pauses.

• 3

  Start long-running speech in the background while continuing other MCP interactions or tasks.

Description

This MCP server wraps the macOS `say` command to provide programmatic text-to-speech via MCP. It offers a `speak` tool with extensive in-text controls (volume, rate, pitch, pauses, emphasis) and a `list_voices` tool to enumerate system voices. The server supports background speech so agents can perform other actions while audio plays and includes examples for integrating with search, transcripts, and notes. Requires macOS and Node.js (>=14).