Hi all,
This is my first post here. I recently built a small open-source tool called audacity-mcp that I thought might interest this community. I didn’t see anything like it out there and I thought it might be useful to somebody.
It’s a Model Context Protocol server that lets Claude (Anthropic’s AI assistant) control Audacity through natural language using the mod-script-pipe interface that Audacity already exposes. No new internals touched.
You can say things like:
-
“Normalize all tracks to -1 dB, then export as FLAC”
-
“Select from 2 to 8 seconds and apply fade in”
-
“Add a label at 12.5 seconds called ‘bridge’”
-
“Apply my CleanPodcast macro to the current selection”
It supports normalize, amplify, fade in/out, noise reduction (two-pass), macros, export/import (WAV/FLAC/MP3/AIFF/OGG), labels, and transport controls. Works on macOS, Linux, and Windows.
The code is MIT-licensed and on GitHub: https://github.com/tachyonshuggy/audacity-mcp
It’s newly released and I’ve tested it against the mod-script-pipe spec, but not across every Audacity version.
Feedback and issue reports are very welcome. Happy to answer questions here or on the repo.
- tachyonshuggy