saitrogen-edge-tts-mcp 0.1.1

MCP server for Microsoft Edge TTS — Remotion-optimised with word-timing JSON

edge-tts MCP Server

Wraps rany2/edge-tts as an MCP server.
No API key. No Windows. No Edge browser. Just Microsoft's TTS — free.

Built for Remotion video production: outputs MP3 + word-timing JSON.

---

Install & Use

The easiest way to use this in your MCP clients (like Cursor or VS Code) is via uvx:

{
  "mcpServers": {
    "edge-tts": {
      "command": "uvx",
      "args": [
        "--from",
        "saitrogen-edge-tts-mcp",
        "edge-tts-mcp"
      ]
    }
  }
}

Or you can install it using pip:

pip install saitrogen-edge-tts-mcp

---

Tools

edgetts_synthesize

Converts text → .mp3 + optional word-timing .json (Remotion-ready).

Param	Default	Description
text	required	Text or SSML to speak
output_path	required	Path ending in .mp3
voice	en-US-AriaNeural	Voice name (see list_voices)
rate	+0%	Speed: +20% faster, -10% slower
volume	+0%	Volume adjustment
pitch	+0Hz	Pitch: +50Hz, -20Hz
save_timings	true	Save <output>.mp3.json with word timings

Word timing JSON (saved at voiceover.mp3.json):

[
  { "word": "Hello",   "startMs": 0,   "durationMs": 312, "endMs": 312 },
  { "word": "Remotion","startMs": 375, "durationMs": 480, "endMs": 855 }
]

edgetts_list_voices

Lists all available voices with optional locale/gender filter.

Param	Example
locale	en-US, hi-IN, ta
gender	Male, Female

---

Remotion Usage

import timings from './public/voiceover.mp3.json';
import { Audio, Sequence, staticFile, useVideoConfig } from 'remotion';

export const Scene = () => {
  const { fps } = useVideoConfig();

  return (
    <>
      <Audio src={staticFile('voiceover.mp3')} />
      {timings.map((w, i) => (
        <Sequence
          key={i}
          from={Math.floor((w.startMs / 1000) * fps)}
          durationInFrames={Math.ceil((w.durationMs / 1000) * fps)}
        >
          <div style={{ color: 'white', fontSize: 48 }}>{w.word}</div>
        </Sequence>
      ))}
    </>
  );
};

---

VS Code / Claude Desktop Config

Add to mcp_servers (stdio transport):

{
  "edge-tts": {
    "command": "python3",
    "args": ["/path/to/edge-tts-mcp/server.py"]
  }
}

Or with uv run:

{
  "edge-tts": {
    "command": "uv",
    "args": ["run", "--with", "edge-tts", "--with", "mcp", "python3", "server.py"]
  }
}

---

Popular Voices

Voice	Language	Gender
en-US-AriaNeural	English (US)	Female
en-US-GuyNeural	English (US)	Male
en-GB-SoniaNeural	English (UK)	Female
hi-IN-SwaraNeural	Hindi	Female
hi-IN-MadhurNeural	Hindi	Male
ta-IN-PallaviNeural	Tamil	Female
ml-IN-SobhanaNeural	Malayalam	Female

Run edgetts_list_voices with locale: "ml" to find Malayalam voices.