Content
# Audacity MCP Server
MCP (Model Context Protocol) server for controlling Audacity through `mod-script-pipe`.
This server communicates with Audacity over named pipes and exposes many Audacity scripting commands as MCP tools.
## Features
- Connects to Audacity via `mod-script-pipe`
- Exposes a broad set of Audacity commands as MCP tools
- Uses `stdio` transport (works with Codex/Claude/Desktop clients)
- Auto-detects pipe suffixes in `/tmp` (for example `.501`)
## Requirements
- Audacity 3.x+
- Python 3.10+
- `mod-script-pipe` enabled in Audacity
## Install
```bash
git clone https://github.com/An-3/mcp-audacity.git
cd mcp-audacity
./scripts/setup.sh
```
## Quick local test (no MCP host needed)
```bash
# 1) Start Audacity and enable mod-script-pipe (see section below)
# 2) Verify local connectivity:
./scripts/smoke_test.sh
# 3) Start MCP server:
./scripts/run.sh
```
If you prefer manual setup instead of scripts:
```bash
python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install "mcp[cli]>=1.6.0" "httpx>=0.28.1"
```
Optional (install as package + CLI entrypoint):
```bash
pip install .
audacity_mcp_server
```
## Enable mod-script-pipe in Audacity
1. Open Audacity.
2. Open the **Modules** preferences page:
- **Windows/Linux**: `Edit -> Preferences -> Modules`
- **macOS**: `Audacity -> Settings... -> Modules` (or `Audacity -> Preferences -> Modules` depending on build)
3. Set `mod-script-pipe` from `New` (or `Disabled`) to `Enabled`.
4. Restart Audacity.
Note: In current Audacity versions, this is configured in **Modules**, not in a separate "Scripting/Remote Control" section.
Verify pipes exist:
```bash
ls -l /tmp/audacity_script_pipe.to.* /tmp/audacity_script_pipe.from.*
```
If the files are present, Audacity is ready.
## Run the server manually
```bash
./scripts/run.sh
```
Expected startup logs include:
- `Audacity MCP server starting up`
- `Opened Audacity mod-script-pipe`
- `Connected to Audacity mod-script-pipe`
## Connect from Codex
Add server to Codex MCP config:
```bash
codex mcp add AudacityMCP -- /absolute/path/to/mcp-audacity/.venv/bin/python /absolute/path/to/mcp-audacity/audacity_mcp_server.py
```
List configured MCP servers:
```bash
codex mcp list
```
## Connect from Claude Desktop
Example `claude_desktop_config.json`:
```json
{
"mcpServers": {
"audacity": {
"command": "/absolute/path/to/mcp-audacity/.venv/bin/python",
"args": [
"/absolute/path/to/mcp-audacity/audacity_mcp_server.py"
]
}
}
}
```
## Troubleshooting
- `Failed to open Audacity pipes`:
- Confirm Audacity is running.
- Confirm `mod-script-pipe` is `Enabled`.
- Confirm pipe files exist in `/tmp`.
- `spawn uv ENOENT`:
- Use absolute paths in client config, or use the Python command shown above instead of `uv`.
- Server starts but tools behave inconsistently:
- Ensure you are using this latest version, which sends commands in `Command:` form and reads full responses until `BatchCommand finished`.
## License
MIT
Connection Info
You Might Also Like
markitdown
Python tool for converting files and office documents to Markdown.
OpenAI Whisper
OpenAI Whisper MCP Server - 基于本地 Whisper CLI 的离线语音识别与翻译,无需 API Key,支持...
oh-my-opencode
Background agents · Curated agents like oracle, librarians, frontend...
claude-flow
Claude-Flow v2.7.0 is an enterprise AI orchestration platform.
chatbox
User-friendly Desktop Client App for AI Models/LLMs (GPT, Claude, Gemini, Ollama...)
continue
Continue is an open-source project for seamless server management.