say-mcp-server

What is say-mcp-server?

say-mcp-server is an MCP server that provides text-to-speech functionality using macOS’s built-in say command. It enables AI assistants to speak text aloud with extensive customization, intended for macOS users who want auditory output from their MCP tools.

How to use say-mcp-server?

Install the package via npm install say-mcp-server, then add the server to your MCP settings configuration file. Once configured, use the speak tool to read text aloud or the list_voices tool to see available system voices.

Key features of say-mcp-server

• Uses macOS’s built-in say command for text-to-speech
• Supports voice selection, speaking rate, and volume control
• Allows speech to run in the background
• Includes advanced tags for pitch, emphasis, and silence insertion
• Lists all available system voices with one command

Use cases of say-mcp-server

• Read search results aloud using Marginalia Search
• Narrate YouTube video transcripts in real time
• Provide audio feedback from Apple Notes content
• Run long speech in background while performing other MCP actions

FAQ from say-mcp-server

What does say-mcp-server do?

It gives an MCP client the ability to convert text into spoken audio on macOS using the system’s say command.

What are the requirements to run say-mcp-server?

You need macOS (for the built-in say command) and Node.js version 14.0.0 or higher.

How can I customize the speech output?

You can set the voice, speaking rate (1–500 words per minute), and use inline tags like [[slnc 500]] for silence, [[rate 200]] for speed changes, [[volm 0.5]] for volume, and [[emph +]] for emphasis.

Can the speech run without blocking other MCP interactions?

Yes, by setting the background parameter to true the speech runs asynchronously, allowing you to perform other MCP tool calls immediately after.

Does say-mcp-server work on Windows or Linux?

No. It relies on macOS’s say command and will not function on non-macOS systems.