ableton-mcp-extended
Enables natural language control of Ableton Live through Claude, including session management, track and clip creation, device parameter adjustment, and browser browsing.
README
Ableton MCP Extended
Complete Claude AI integration for Ableton Live through Model Context Protocol (MCP)
Control Ableton Live directly from Claude with natural language - create tracks, compose melodies, mix songs, and produce music using AI assistance!
🎵 Features
Session Control
- Playback Control: Start/stop playback, control recording
- Tempo Management: Set and adjust session tempo
- Session Info: Get real-time session status
Track Management
- Create Tracks: MIDI and audio tracks with custom names
- Track Info: Get detailed track information including clips and devices
- Track Control: Volume, mute, solo, and track naming
- Track Deletion: Remove unwanted tracks
Clip Operations
- Create Clips: MIDI clips with custom length and names
- Add Notes: Programmatic MIDI note creation with pitch, timing, velocity
- Clip Control: Fire, stop, and delete clips
- Clip Naming: Organize clips with descriptive names
Device & Effects
- Device Info: Get information about track devices and parameters
- Parameter Control: Adjust device parameters programmatically
- Instrument Loading: Load instruments and effects from browser
Browser Integration
- Browse Content: Explore Ableton's browser hierarchy
- Load Instruments: Load instruments, drum kits, and effects
- Sample Management: Access and load samples and packs
🚀 Quick Start
1. Installation
Run the automated installer:
cd ableton-mcp-extended
python install_remote_script.py
The installer will:
- 🔍 Auto-detect your Ableton Live installation
- 📦 Install the remote script
- ⚙️ Configure Claude Desktop
- 🎯 Set up the Python package
2. Ableton Live Setup
- Start Ableton Live
- Go to Preferences → MIDI
- Control Surface: Select
AbletonMCP_Extended_Remote_Script - Input/Output: Leave as "None"
3. Claude Desktop Setup
Restart Claude Desktop after installation. The MCP server should appear in your configuration.
4. Start Creating Music!
Open Claude and try commands like:
"Create a drum track and add a basic beat"
"Add a bass line in C minor"
"Create a chord progression using piano"
"Set the tempo to 128 BPM and start playback"
🎹 Usage Examples
Creating a Basic Song Structure
Hey Claude, let's create a song:
1. Set the tempo to 120 BPM
2. Create a drum track with a basic 4/4 beat
3. Add a bass track with a simple bass line in C major
4. Create a piano track with a chord progression: C - Am - F - G
5. Start playback so I can hear it
Detailed Music Production
I want to create a house track:
1. Set tempo to 128 BPM
2. Create a drum track called "House Drums"
3. Add a classic house beat with:
- Kick on beats 1 and 3
- Snare on beats 2 and 4
- Hi-hats on every 8th note
4. Create a bass track and add a pumping bass line
5. Add some reverb to the drums and adjust the mix
Live Performance Control
During my live set:
1. Mute the bass track
2. Gradually increase the volume of track 3
3. Fire the clip in slot 2 of the drum track
4. Stop all clips and restart from the beginning
🔧 Manual Installation
If the automatic installer doesn't work, follow these steps:
1. Install Python Package
pip install -e .
2. Install Remote Script
Copy AbletonMCP_Extended_Remote_Script to your Ableton Live MIDI Remote Scripts directory:
macOS:
/Applications/Ableton Live [Version].app/Contents/App-Resources/MIDI Remote Scripts/
Windows:
C:\Program Files\Ableton\Live [Version]\Resources\MIDI Remote Scripts\
3. Configure Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"ableton-mcp-extended": {
"command": "uvx",
"args": ["ableton-mcp-extended"]
}
}
}
Config file locations:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
~/AppData/Roaming/Claude/claude_desktop_config.json - Linux:
~/.config/claude/claude_desktop_config.json
🧪 Testing
Run the comprehensive test suite:
# Install test dependencies
pip install pytest pytest-asyncio pytest-timeout
# Run all tests
pytest
# Run specific test categories
pytest tests/test_ableton_connection.py # Connection tests
pytest tests/test_music_creation.py # Music creation workflows
pytest tests/test_integration.py # Integration tests
🎛️ Available Commands
Session Commands
get_session_info()- Get tempo, time signature, playback stateset_tempo(tempo)- Set session tempo (20-999 BPM)start_playback()- Start playbackstop_playback()- Stop playbackset_session_record(enabled)- Enable/disable recording
Track Commands
create_midi_track(name, position)- Create MIDI trackcreate_audio_track(name, position)- Create audio trackget_track_info(track_index)- Get track detailsset_track_name(track_index, name)- Rename trackset_track_volume(track_index, volume)- Set volume (0.0-1.0)set_track_mute(track_index, mute)- Mute/unmute trackset_track_solo(track_index, solo)- Solo/unsolo trackdelete_track(track_index)- Delete track
Clip Commands
create_clip(track_index, slot_index, length_bars, name)- Create clipadd_notes_to_clip(track_index, slot_index, notes)- Add MIDI notesset_clip_name(track_index, slot_index, name)- Rename clipfire_clip(track_index, slot_index)- Play clipstop_clip(track_index, slot_index)- Stop clipdelete_clip(track_index, slot_index)- Delete clip
Browser Commands
get_browser_tree()- Get browser structureget_browser_items(path)- Get items at pathload_browser_item(item_path, track_index)- Load itemload_instrument(name, track_index)- Load instrumentload_drum_kit(name, track_index)- Load drum kit
Device Commands
get_device_info(track_index, device_index)- Get device infoset_device_parameter(track_index, device_index, param_index, value)- Set parameter
🎼 MIDI Note Format
When adding notes to clips, use this format:
notes = [
{
'pitch': 60, # MIDI note number (60 = C4)
'time': 0.0, # Start time in beats
'duration': 1.0, # Note length in beats
'velocity': 100, # Velocity (1-127)
'mute': False # Optional: mute state
}
]
Common MIDI Notes:
- C4 = 60, C#4 = 61, D4 = 62, D#4 = 63, E4 = 64
- F4 = 65, F#4 = 66, G4 = 67, G#4 = 68, A4 = 69
- A#4 = 70, B4 = 71, C5 = 72
Drum Notes (General MIDI):
- Kick = 36, Snare = 38, Hi-Hat = 42, Open Hi-Hat = 46
- Crash = 49, Ride = 51
🔧 Troubleshooting
Connection Issues
"Failed to connect to Ableton"
- Ensure Ableton Live is running
- Check that the remote script is properly installed
- Verify the remote script is selected in MIDI preferences
- Try restarting Ableton Live
"Remote script not appearing in preferences"
- Check the script was copied to the correct directory
- Ensure the
__init__.pyfile is present - Try refreshing or restarting Ableton Live
Claude Integration Issues
"MCP server not found"
- Restart Claude Desktop after configuration
- Check the config file syntax and location
- Verify the Python package is installed:
pip list | grep ableton-mcp
"Commands not working"
- Check connection status: Ask Claude "What's the Ableton connection status?"
- Verify Ableton is running and the remote script is active
- Check for error messages in Claude
Performance Issues
"Slow response times"
- Reduce the number of concurrent operations
- Use smaller clip lengths for testing
- Check system resources and close other applications
"Memory usage"
- Restart Ableton Live periodically during long sessions
- Clear unused clips and tracks
- Monitor system memory usage
🤝 Contributing
Contributions are welcome! Please:
- Fork the repository
- Create a feature branch
- Add tests for new functionality
- Ensure all tests pass
- Submit a pull request
📝 License
MIT License - see LICENSE file for details.
🙏 Acknowledgments
- Built on the Model Context Protocol (MCP) by Anthropic
- Inspired by the Ableton Live community
- Uses FastMCP for efficient server implementation
🔗 Links
Start making music with AI today! 🎵✨
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。