AudacityMCP
MCP server for Audacity 3.x with 144 tools and 9 pipelines for audio cleanup, mastering and transcription, all processed locally.
- 131 ★ on the repository
- Last commit
Extended access
Why this level
- The agent edits and exports Audacity project files, and the installer changes the Claude Desktop configuration and enables the scripting module.
Install
- Works in
Manual install
git clone https://github.com/xDarkzx/Audacity-MCP.git
cd Audacity-MCP
bash install.shThe installer installs the package from the local folder, enables `mod-script-pipe` and configures Claude Desktop. On Windows run `install.bat`.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent edits and exports Audacity project files, and the installer changes the Claude Desktop configuration and enables the scripting module.
This is third-party code. Review the repository files before installing.
What it does
AudacityMCP controls an open Audacity through `mod-script-pipe`: effects, editing, tracks, labels, import and export, analysis. The agent builds chains itself or runs one of nine ready pipelines: podcast, interview, vocal and live recording cleanup, music mastering with six genre presets, audiobook preparation for ACX, and a lo-fi effect. Pipelines run in the background and return a `job_id`, and you check status with `check_pipeline_status`. They only reduce peaks and never raise loudness, and the target LUFS level is set separately with `loudness_normalize`. Transcription through faster-whisper is experimental, runs offline, places labels and saves SRT and VTT subtitles.
Who it is for. Podcasters, musicians and video creators who clean and master recordings in Audacity with agent commands.
Good fit when
- A podcast with background noise: noise profile, high-pass filter and compression in one pipeline.
- A track is prepared for streaming with a genre preset, then set to -14 LUFS.
- An interview is transcribed offline and labels are placed on every phrase.
- A recording holds several takes: labels mark the extras, and the agent cuts them and exports the segments as separate files.
Not a fit when
- You use Audacity 4: it has no `mod-script-pipe`, and a separate Audacity4-MCP project exists for it.
- You need processing without an open Audacity: the server does not launch the program itself.
- The original recording exists only in the project with no backup.
Example request
Clean up this podcast recording, then normalize it to -16 LUFS and export as FLAC.MCP
- Transport
- stdio
- Authentication
- not required
Limitations
It works only with Audacity 3.x, because Audacity 4 has no `mod-script-pipe`. Audacity must be running first with the module enabled in the Modules preferences, then restarted. Python 3.10 or later is required. Transcription is experimental and needs separate setup, and GPU acceleration works only with NVIDIA cards. The client may lack the PATH for the `audacity-mcp` command, in which case set the full path to the executable. Long effects on long files take minutes and produce channel timeouts while Audacity is busy.
How to disable. Remove the `audacity` entry from the client configuration and run `pip uninstall audacity-mcp-server`. Disable the `mod-script-pipe` module in Audacity preferences separately.
FAQ
How do I install AudacityMCP?
Clone the repository and run `bash install.sh` (`install.bat` on Windows): the installer enables `mod-script-pipe` and, after confirmation, configures Claude Desktop. For other clients use `pip install audacity-mcp-server` and an entry with the `audacity-mcp` command.
Does AudacityMCP work with Audacity 4?
No. A separate Audacity4-MCP project exists for Audacity 4 and it needs an Audacity fork built from source.
Why do pipelines leave loudness alone?
They only reduce peaks that are too hot so a bad recording is not blown out. Set the target level separately with `loudness_normalize`, for example `lufs_level=-14`.
Does audio leave the machine?
No. Processing runs locally through the Audacity pipe, and faster-whisper transcription also runs offline.
How do I enable transcription?
Transcription is experimental and needs separate setup from the Transcription Setup section of `docs/INSTALLATION.md`. For NVIDIA acceleration run `audacity-mcp-setup-gpu`.
Discussion
No questions or reviews yet.
Related
MCP serversA local MCP server generates music and sound effects, builds loops, adjusts loudness and exports audio for games or video.
A local sound studio with an MCP server: an agent makes effects, music, songs and speech on open-weight models
An MCP server for Ableton Live: the agent creates tracks, places clips and builds arrangements straight in the open project
MCP server for REAPER: an agent creates projects, tracks and MIDI, adds effects, mixes, masters and renders a track
SmartCut MCP
SmartCut MCP Server
MCP server that edits talking head videos inside a CapCut project: it uses auto-captions to cut silences and repeated takes.
An MCP server for Adobe Photoshop: background removal, retouch, social export and batch jobs by chat
README in short
The 144 tools fall into 11 categories: effects, cleanup, editing, project, tracks, selection, transport, analysis, generation, transcription and labels. The installer supports a `--dry-run` mode and backs up configuration before changing it.