# MCP Listen

> MCP Listen is an open-source MCP server for ai models and llm tooling. Microphone capture and local speech-to-text for MCP agents, with optional Ollama replies. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** Windows, macOS, Linux, CLI
- **Repository:** https://github.com/decibri/mcp-listen
- **MCP Registry name:** io.github.decibri/mcp-listen
- **Categories:** [AI models and LLM tooling](https://ai-powered.com/categories/ai-and-llm), [design and media](https://ai-powered.com/categories/design-and-media)
- **Listing updated:** 2026-11-18

## About

Three tools give an agent ears. It can list microphone devices, record a WAV file, or run a full voice query: capture speech, transcribe it on the same machine with whisper.cpp, then send the text to a model running in a local Ollama daemon and return both transcript and reply. The README states the audio stays on the machine.

Recording can stop on a fixed duration (100 ms to 30 seconds) or when the speaker finishes, using a bundled Silero voice-activity detector that needs no download. Results report how a recording ended, such as silence, ceiling or no speech after ten seconds, and the voice query separates silence, speech without words and real errors so callers can branch on fields rather than prose. Arguments are validated and unknown ones rejected before anything is written.

It works on Windows x64, Apple silicon Macs and Linux x64 or arm64, but not Intel Macs. The voice query additionally needs Ollama and a Whisper GGML model.

## Features

- **Device listing:** Returns input devices with stable ids, channels and sample rates.
- **WAV capture:** Records for a set time or until the speaker stops, saving a 16 kHz WAV.
- **Silence detection:** Uses on-device Silero voice-activity detection to end recordings.
- **Local transcription:** Transcribes with whisper.cpp without sending audio away.
- **Ollama voice query:** Passes the transcript to a local model and returns its reply.
- **Strict arguments:** Rejects unknown or invalid arguments before anything is recorded.

## Good for

- Let a coding assistant take spoken instructions
- Record a short voice note to a WAV file
- Ask a local model a question by voice
- Pick a specific microphone for an agent

## Install

### Claude Code (Terminal)

```sh
claude mcp add mcp-listen -- npx -y mcp-listen
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "mcp-listen": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-listen"
      ]
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "mcp-listen": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-listen"
      ]
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "mcp-listen": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "mcp-listen"
      ]
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "mcp-listen": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-listen"
      ]
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "mcp-listen": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-listen"
      ]
    }
  }
}
```

## Tools

- `list_audio_devices`: Lists available microphone input devices
- `capture_audio`: Records microphone audio and saves it as a WAV file
- `voice_query`: Captures speech, transcribes locally and queries a local Ollama model

## FAQ

### Does audio leave my machine?

For voice_query the README says capture, transcription and the model call all stay local.

### Which platforms work?

Windows x64, macOS Apple silicon, and Linux x64 and arm64; Intel Macs are not supported.

### What do I need for voice queries?

Ollama running locally and a Whisper GGML model; Node.js 18 or newer and a microphone for basic capture.

### How do I install it?

Run claude mcp add mcp-listen npx mcp-listen, or add npx -y mcp-listen to your client's MCP config.

### Is MCP Listen free?

Yes. MCP Listen is open source, so you can use and self-host it for free.

### How do I add MCP Listen to Claude Code?

Run this command in your terminal: claude mcp add mcp-listen -- npx -y mcp-listen

### What are the best alternatives to MCP Listen?

The closest alternatives to MCP Listen are HyperFrames by HeyGen, headroom, omniget, next-ai-draw-io and mcp-for-blender.

## Where this comes from

- Description: written from the project's README.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [HyperFrames by HeyGen](https://ai-powered.com/mcp/hyperframes-by-heygen): HyperFrames by HeyGen is a freemium MCP server for design and media.
- [headroom](https://ai-powered.com/mcp/headroom-3): headroom is an open-source MCP server for ai models and llm tooling.
- [omniget](https://ai-powered.com/mcp/omniget): omniget is an open-source MCP server for ai models and llm tooling.
- [next-ai-draw-io](https://ai-powered.com/mcp/next-ai-draw-io-2): next-ai-draw-io is an open-source MCP server for ai models and llm tooling.
- [mcp-for-blender](https://ai-powered.com/mcp/mcp-for-blender): mcp-for-blender is an open-source MCP server for design and media.
- [ffmpeg-skill](https://ai-powered.com/mcp/ffmpeg-skill): ffmpeg-skill is an open-source MCP server for design and media.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [unity-mcp](https://ai-powered.com/mcp/unity-mcp-2): unity-mcp is an open-source MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/mcp-listen
