MCP-Say: Instant Ultra-Realistic Voices

MCP-Say: Transform text into lifelike voices instantly! Ultra-realistic TTS server delivers crisp, customizable audio—powering apps, games, and AI with flawless, human-like speech." )

Visit Repository

✨ Developer Tools

4.8(133 reviews)

199 saves

93 comments

This tool saved users approximately 13055 hours last month!

About MCP-Say

What is MCP-Say: Instant Ultra-Realistic Voices?

Ever wanted your AI tools to speak with the clarity of a real human? MCP-Say bridges that gap by turning text into lifelike audio. This lightweight TTS server plugs directly into platforms like Claude Desktop and Cursor IDE, giving developers instant voice capabilities without reinventing the wheel. Under the hood, it harnesses both macOS’s native ‘say’ command and ElevenLabs’ advanced API—so you get simplicity paired with enterprise-grade realism.

How to Use MCP-Say: Instant Ultra-Realistic Voices?

Getting started is as smooth as a well-rehearsed monologue. First, install the server with a single Go command. Then configure your IDE to point at MCP-Say’s endpoint—don’t forget to secure your ElevenLabs API key in the env block. Test it live by piping JSON requests through the CLI. Within minutes, your code reviews will be narrated, and error logs will have personality.

MCP-Say Features

Key Features of MCP-Say: Instant Ultra-Realistic Voices?

Dual Engine Flexibility: Toggle between macOS’s built-in TTS for speed or ElevenLabs’ hyper-realistic voices for polished outputs
MCP Protocol Native: Designed from the ground up to work with modern developer tool ecosystems
Log-Level Control: Debug mode gives granular visibility into speech synthesis workflows
Zero-Setup Base: Leverages existing OS capabilities while offering enterprise API scalability

Use Cases of MCP-Say: Instant Ultra-Realistic Voices?

Perfect for developers building voice-enabled apps, but shines brightest in these scenarios:

Automating accessibility features in IDEs
Prototyping conversational interfaces with realistic audio
Creating vocal feedback loops during code debugging
Generating training materials with synthetic voices

MCP-Say FAQ

FAQ from MCP-Say: Instant Ultra-Realistic Voices?

Does it work on Windows? The ‘say’ tool is macOS-only, but ElevenLabs API support works cross-platform

Can I customize voice speeds? Yes! macOS ‘say’ accepts rate parameters through the API payload

Is it free? The server is MIT-licensed, but ElevenLabs requires a paid plan for their voice models

How stable is the API? Battle-tested with daily CI/CD pipelines—check those passing GitHub Actions badges!

Content

mcp-say

TTS (text-to-speech) MCP Server

What? 🤔

Adds Text-to-Speech to things like Claude Desktop and Cursor IDE.

It registers two tools:

say
elevenlabs

`say`

Uses the macOS say binary to speak the text

`elevenlabs`

Uses the elevenlabs text-to-speech API to speak the text

Getting Started

Install

go install github.com/blacktop/mcp-say@latest



❱ mcp-say --help

TTS (text-to-speech) MCP Server.

Provides a text-to-speech service using the MacOS 'say' command.

Designed to be used with the MCP protocol.

Usage:
  mcp-say [flags]

Flags:
  -h, --help      help for mcp-say
  -v, --verbose   Enable verbose debug logging

Set Claude Desktop Config

{
  "mcpServers": {
    "say": {
      "command": "mcp-say",
      "env": {
        "ELEVENLABS_API_KEY": "********"
      }
    }
  }
}

Test

❱ cat test/say.json | go run main.go --verbose

2025/03/23 22:41:49 INFO Starting MCP server name="Say TTS Service" version=1.0.0
2025/03/23 22:41:49 DEBU Say tool called request="{Request:{Method:tools/call Params:{Meta:<nil>}} Params:{Name:say Arguments:map[text:Hello, world!] Meta:<nil>}}"
2025/03/23 22:41:49 DEBU Executing say command args="[--rate 200 Hello, world!]"
2025/03/23 22:41:49 INFO Speaking text text="Hello, world!"



{"jsonrpc":"2.0","id":3,"result":{"content":[{"type":"text","text":"Speaking: Hello, world!"}]}}