# Audio Intelligence MCP server

> Enables AI agents to perform local-first audio analysis including transcription, speaker diarization, sound event detection, and grounded question answering.

- Repository: https://github.com/DNSdecoded/audio-intelligence-mcp
- Page: https://hostedmcp.org/servers/dnsdecoded/audio-intelligence-mcp
- Transport: —
- Stars: 2
- License: Apache-2.0

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `AUDIO_MCP_ALLOWED_ROOTS` | no | no |  |

## Tools (13)

- `analyze_audio`
- `transcribe_audio`
- `detect_audio_events`
- `extract_audio_features`
- `inspect_audio_quality`
- `get_audio_timeline`
- `ask_audio`
- `summarize_audio`
- `compare_audio`
- `search_audio_library`
- `analyze_audio_stream`
- `export_audio_analysis`
- `get_job`

## FAQ

### Is my audio data private?

Yes, everything runs on your machine and audio never leaves it unless you explicitly enable an LLM provider.

### What models are used for transcription?

The server uses faster-whisper for transcription tasks.

### Does it require a GPU?

A CPU is sufficient, but CUDA is automatically used if a compatible GPU is available.
