# Audio MCP [Health: Active]

**Category:** 🎥 Multimedia Process  
**Repository:** https://github.com/mshark1/audio-mcp  
**GitHub Stars:** 0  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/audio-mcp

## Description
Audio perception for AI agents: measure, see, compare, edit and synthesize sound locally.

## Tools
Capabilities this server exposes over MCP:

- **audio_analyze** — Peak, true peak, RMS, LUFS loudness, clipping, DC offset, a loudness sparkline, silence and noise floor, stereo correlation/width/balance, band energy, brightness, tonal vs. noisy, strongest frequencies as note names, pitch, onsets, tempo, edge clicks, loop-seam quality, and an **`issues`** list in…
- **audio_events** — Splits audio into distinct sounds and describes each one: timing, level, attack, tonal/noisy, brightness, pitch or glide
- **audio_view** — PNG waveform and/or spectrogram, with an optional pitch overlay. Every column covers its full time span, so short clicks aren't skipped on long files
- **audio_timeline** — Level and brightness over time. Streams the file, so it works on recordings of any length
- **audio_melody** — Single-line transcription: notes, timings, cents, compact text
- **audio_transcribe** — Speech to text with local Whisper: timed words, language, speaking rate, long pauses and **uncertain words** *(optional extra)*
- **audio_compare** — A/B summary, metric deltas, per-band dB changes, issues fixed or introduced, and a **null test** (identical? only a gain change? lightly changed?)
- **audio_process** — An ordered chain of edits written to a **new** file: gain, peak/LUFS normalize, true-peak limiter, compressor, noise reduction, RMS normalize, trim, fades, crop, pad, high/low-pass, peaking EQ, shelves, DC removal, polarity, mono/stereo, resample, reverse, seamless-loop crossfade, mix, append
- **audio_batch_survey** — Audits a folder for consistency (loudness outliers, mixed sample rates/channels/formats, uneven leading silence) and returns ready-to-run fixes
- **audio_batch_process** — Applies one edit chain to many files and writes them to an output folder
- **audio_synthesize** — Declarative synth: band-limited oscillators with phase control, noise colours (pink/brown loop seamlessly), sweeps, ADSR, vibrato, tremolo, filters, pan, note sequences, plus exact test signals (impulse, DC, silence)
- **audio_info** — Metadata for a file, or a listing of the audio files in a folder
- **audio_check_delivery** — Pass/fail against a delivery spec** (Apple Podcasts, Spotify, YouTube, EBU R 128, ATSC A/85, ACX audiobooks, game SFX) with cited sources and a fix that has already been tried on your file
- **audio_check_game_project** — Audits a Godot 4 project's audio**: loudness consistency per category, loop settings and clicky seams, defects, and broken or unused `res://` references

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "audio-mcp": {
    "command": "uvx",
    "args": ["audio-mcp"],
    "env": {
      "AUDIO_MCP_ALLOWED_DIRS": "",
      "AUDIO_MCP_MAX_SECONDS": "",
      "AUDIO_MCP_WHISPER_MODEL": "",
      "AUDIO_MCP_ROOT": "",
      "AUDIO_MCP_FFMPEG": ""
    }
  }
}
```

**Requires environment variables:** `AUDIO_MCP_ALLOWED_DIRS`, `AUDIO_MCP_MAX_SECONDS`, `AUDIO_MCP_WHISPER_MODEL`, `AUDIO_MCP_ROOT`, `AUDIO_MCP_FFMPEG` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What Audio MCP does

The Audio MCP MCP server exposes local audio workflows through MCP for the 🎥 Multimedia Process category. Agents can inspect files and folders, measure loudness and signal characteristics, generate waveform or spectrogram images, identify events, transcribe speech, recover single-line melodies, and compare two versions of a file.

Analysis includes peak and true-peak levels, RMS, LUFS, clipping, DC offset, silence, noise floor, stereo properties, band energy, brightness, tonal or noisy classification, prominent frequencies, pitch, onsets, tempo, edge clicks, and loop-seam quality. Results can include plain-language issues with suggested processing steps.

The Audio MCP MCP server also supports ordered edits that write to a new file by default. Processing operations include normalization, limiting, compression, noise reduction, trimming, fades, filtering, EQ, channel changes, resampling, polarity changes, reversing, looping crossfades, mixing, and appending. Batch tools survey folders for inconsistencies and apply one edit chain to multiple files.

## How it works

The server runs locally and communicates with an MCP client over its standard command configuration. Core dependencies include NumPy, SciPy, SoundFile, and Matplotlib. Speech transcription is an optional installation using local Whisper support; its model downloads on first use.

Audio can be read from WAV, FLAC, OGG/Vorbis, Opus, MP3, AIFF, and CAF through libsndfile. Installing FFmpeg adds support for M4A/AAC, WMA, and audio tracks from MP4, MOV, WebM, and MKV files. Analysis is chunked, and the timeline tool streams input so it can handle recordings of different lengths.

Synthesis uses declarative descriptions for oscillators, colored noise, sweeps, envelopes, modulation, filters, panning, note sequences, and exact test signals. The server supports a verify-after-edit workflow: create or process a sound, analyze it, view it, and compare the result with the source.

## Setup and configuration

Install the package with `pip install audio-mcp` or `uv tool install audio-mcp`. Add the optional speech extra with `pip install "audio-mcp[speech]"`. A typical MCP configuration launches `uvx audio-mcp`; Claude Code can add the server with `claude mcp add audio -- uvx audio-mcp`.

`AUDIO_MCP_ALLOWED_DIRS` restricts readable and writable paths. `AUDIO_MCP_MAX_SECONDS` sets the largest section loaded into memory, while `AUDIO_MCP_WHISPER_MODEL` selects the Whisper model and defaults to `base`. `AUDIO_MCP_ROOT` supplies the base directory for relative paths. Set `AUDIO_MCP_FFMPEG` when FFmpeg is not available on the system path.

## Tools and capabilities

The Audio MCP MCP server provides these tool groups:

- `audio_analyze`, `audio_events`, `audio_view`, and `audio_timeline` measure, classify, and visualize sound.
- `audio_melody` extracts a monophonic note line; `audio_transcribe` returns local Whisper speech results with timing and optional uncertainty information.
- `audio_compare` reports metric changes, per-band differences, issues fixed or introduced, and null-test results.
- `audio_process`, `audio_batch_process`, and `audio_synthesize` create processed or generated files.
- `audio_batch_survey` finds folder-level consistency problems, while `audio_info` reports metadata or lists audio files.
- `audio_check_delivery` tests specifications for services and standards including Apple Podcasts, Spotify, YouTube, EBU R 128, ATSC A/85, ACX, and game sound effects.
- `audio_check_game_project` audits Godot 4 audio, including category loudness, loop settings, seams, defects, and broken or unused `res://` references.

## Limitations and notes

Edits are non-destructive by default and normally produce a processed file. Overwriting requires an explicit option, and batch output cannot target the input folder. Delivery checks can identify fixes that cannot be automated, including cases where heavy background noise needs human judgment. Godot project audits do not modify the project.

Speech transcription requires the optional dependency and downloads a model. FFmpeg is required for the additional container and video formats described above. File access is otherwise governed by the operating-system account unless an allowed-directory restriction is configured.

_Full upstream README: https://allmcps.com/mcp/audio-mcp/readme_