Vox MCP Server

Native macOS MCP server for voice I/O — Swift binary, SFSpeechRecognizer + ElevenLabs TTS.

Local serverstdioPython

What is the Vox MCP server?

Native macOS MCP server for voice I/O — Swift binary, SFSpeechRecognizer + ElevenLabs TTS. Exposed over MCP by the vox mcp server, that capability becomes something an assistant can invoke while it works, not something you go and do afterwards.

What it actually does

Built in Swift. No Node.js. No Python. No Electron. Just a single binary.

  • AVFoundation — — native mic capture, zero overhead
  • SFSpeechRecognizer — — Apple on-device speech recognition, works offline, 50+ languages
  • NLLanguageRecognizer — — automatic language detection per utterance
  • AVSpeechSynthesizer — — built-in TTS fallback, no API key needed
  • Single ~200KB binary. Zero npm install. Zero Python venv

Its toolset

Everything the assistant can do here goes through one of these:

  • listen — Activates the mic, waits for you to speak, returns transcript + detected language when silence is detected
  • speak — Speaks text aloud — ElevenLabs TTS with automatic fallback to macOS system voice

Adding it to your client

Setup follows the usual MCP pattern — install or clone the server, register it in your client's configuration file, restart the client. The configuration blocks on this page cover the common clients.

Configuration

You will need one environment variable: ELEVENLABS_API_KEY. The server will not start without them, which is usually why the tools fail to appear on a first run. Keep credentials in your client's env block or a secrets manager rather than in a file you might commit.

  • macOS 13+ - Microphone set as default in System Settings → Sound → Input - Microphone + Speech Recognition permission for your terminal app - Optional: ELEVENLABS_API_KEY in ~/.claude/.env for high-quality multilingual TTS

When to reach for it

Among the AI and media services options, the useful question is rarely "what can it do" but "what does it cost you to run" — permissions, credentials, and how much of your context its toolset consumes. Vox's toolset — listen, speak — is a fair guide to whether it matches your workflow. It is maintained by boska; worth a glance at recent repository activity before you build anything load-bearing on it.

This entry was verified against Vox's own documentation before publication; SyncDev keeps the directory reviewed rather than auto-generated.

Caveats

  • It runs with your machine's permissions. That is convenient and also the reason to think about what you point it at before you approve a tool call.
  • Missing credentials fail quietly in some clients — if no tools show up, check the environment block first.
  • MCP clients confirm each tool call by default. Leave that on until you have watched what the vox mcp server does with a few real requests.

Available tools

ToolWhat it does
listenActivates the mic, waits for you to speak, returns transcript + detected language when silence is detected
speakSpeaks text aloud — ElevenLabs TTS with automatic fallback to macOS system voice

How to install the Vox MCP server

{
  "mcpServers": {
    "vox": {
      "command": "/path/to/vox"
    }
  }
}

Configuration as documented by the project. Restart the client after saving.

Configuration

  • macOS 13+ - Microphone set as default in System Settings → Sound → Input - Microphone + Speech Recognition permission for your terminal app - Optional: ELEVENLABS_API_KEY in ~/.claude/.env for high-quality multilingual TTS
VariableDescriptionRequired
ELEVENLABS_API_KEYCredential the server authenticates with.Yes

Example prompts to try

  • Use Vox to listen.
  • Use Vox to speak.

Frequently asked questions

It connects Vox to MCP-compatible AI assistants such as Claude and Cursor, exposing 2 tools (listen, speak) that the assistant can call on your behalf. Instead of copying data back and forth by hand, the assistant works with Vox directly.