Provides tools for editing audio using the Audacity 4 audio engine, including splitting, trimming, converting, resampling, applying effects, mixing, joining, and installing an in-app Audacity extension to add audio tools to Audacity's menus.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Audacity MCP Server split every recording in D:\voice into 15-second pieces"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Audacity 4 is a great open-source audio editor, but it has no scripting interface (no mod-script-pipe, no command-line export). This Model Context Protocol server gives Claude access to Audacity's own audio engine, so you can ask things like:
"Split every recording in
D:\voiceinto 15-second pieces, one folder per file." "ConvertD:\podcast\ep1.wavto FLAC 24-bit at 48 kHz." "Apply a high-pass at 80 Hz, a compressor and normalize to -1 dBFS onD:\take3.wav."
Inspired by VectorMagic-mcp-server.
Related MCP server: Audacity MCP Server
Part | What it is |
Native engine ( | Loads Audacity 4's own codec libraries at runtime (libsndfile 1.2.2, mpg123) from its install folder, so it reads and writes exactly what your Audacity supports. Streams long files; high-quality windowed-sinc resampling. |
MCP server (Node.js) | 10 tools for Claude. Optional FFmpeg for MP3/M4A/AAC/WMA export and exotic imports. |
Audacity extension ("MCP Audio Tools") | An official Audacity 4 extension (JavaScript + native DLL) that adds effects to Audacity's own menus. Installed with one tool call. |
Tool | What it does |
| Cut one or many files into N-second pieces (default 15 s), one folder per input, numbered |
| Change format, bit depth, sample rate or channels of one or many files |
| Keep a time range, with optional fade in/out |
| Chain of effects (see below) |
| Mix files on top of each other with per-file gain and start offset |
| Join files with an equal-power crossfade or a silent gap |
| Format, rate, channels, duration, peak and RMS level |
| Open results in the Audacity 4 editor |
| Install / update / remove the in-app extension |
| Audacity path and version, engine, FFmpeg, formats, effects |
Input (native) | wav, aiff, flac, ogg, opus, mp3, mp2, caf, w64, rf64, au, voc, sd2, iff, paf… |
Input (via FFmpeg) | m4a, aac, mp4, wma, wv, ac3, webm, mkv, ape… |
Output (native) | wav, flac, ogg (Vorbis), opus, aiff, caf, w64, rf64, au |
Output (via FFmpeg) | mp3, m4a, aac, wma |
Options | bit depth |
gain · normalize · fade_in · fade_out (linear, exponential, logarithmic, S-curve) · trim · pad · trim_silence · reverse · speed · invert · remove_dc · highpass · lowpass · bandpass · notch · eq · bass · treble · echo · compressor · limiter · noise_gate · mono · stereo · swap_channels · pan · mouth_declick · declip · noise_reduction · click_removal
Example: [{"type":"highpass","frequency":80},{"type":"compressor","threshold_db":-20,"ratio":3},{"type":"normalize","peak_db":-1}]
Voice cleanup / limpieza de voz: [{"type":"declip"},{"type":"mouth_declick","sensitivity":6},{"type":"limiter","ceiling_db":-1}]
After install_audacity_extension and restarting Audacity 4:
Menu | Effect |
Tools › Extension › audacity-mcp-server | Split into Files (MCP): writes the selection (or each selected track) as N-second files |
Export Selection to File (MCP): WAV, FLAC, OGG, Opus, AIFF | |
Import Audio File as Track (MCP): any supported file (including MP3) as a new track | |
Analyze › Extension › audacity-mcp-server | Labels Every N Seconds (MCP): a label track with one region per interval |
Tools | MCP Audio Tools: status: checks the native library |
All of them support undo and show a progress bar.
Windows 10/11 x64
Audacity 4 (tested with 4.0.1)
Node.js ≥ 20 (bundled with Claude Desktop for .mcpb installs)
Optional: FFmpeg (only for MP3/M4A/AAC/WMA export or M4A/AAC/WMA/WavPack import)
Download audacity-mcp-server.mcpb from the latest release.
Double-click it (or drag it onto Claude Desktop → Settings → Extensions) and click Install.
Optional: if Audacity is not in C:\Program Files\Audacity 4, set its folder in the extension settings; set the FFmpeg path if it is not on PATH.
Optional: ask Claude "install the Audacity extension", then restart Audacity.
git clone https://github.com/jhonsu01/audacity-mcp-server.git
cd audacity-mcp-server
npm install
npm run build:all
claude mcp add audacity -- node "%CD%\dist\bundle.cjs"build:all compiles the native parts and needs Visual Studio with the C++ workload.
Variable | Meaning |
| Audacity 4 install folder (or |
|
|
| Max seconds per job (default 1800) |
Files are never overwritten unless overwrite: true.
All paths must be absolute; the input is never modified.
Everything runs locally, with no network access.
npm run build:native # aumcp-engine.exe + audacity_mcp_native.dll (MSVC)
npm run build # TypeScript + bundle
npm run test:e2e # 25 end-to-end checks over real MCP stdio (synthetic audio)
npm run pack:mcpb # audacity-mcp-server.mcpbGPL-3.0-only. The extension ABI header comes from Audacity (GPL-3.0). Audacity® is a trademark of Muse Group. This project is not affiliated with or endorsed by Audacity or Muse Group.
This server cannot be deployed
One workspace of tools for Claude and ChatGPT: connect 600+ apps, generate media, build tools.
Use your Mac, Windows or Linux computer from ChatGPT, Claude or Codex: files, commands, documents.
Audio for your agent: transcribe, speak, translate, summarise, plus sound effects and music.
Generate AI music via the Lacuna Music API from MCP clients like Claude Desktop & Code.
