AbletonMCP (Ableton Live MCP Server)
Author: ahujasid
Description: MCP (Model Context Protocol) server + Ableton Live Remote Script that lets MCP clients (Claude Desktop / Cursor) control Ableton Live via a JSON-over-TCP socket bridge. Enables session/track inspection, track and clip creation/editing, device (instruments/effects) loading from Ableton’s browser, transport/playback control, tempo changes, and MIDI note insertion.
Stars: 2.3k
Forks: 281
License: MIT License
Category: Specialized
Overview
Installation
npx -y @smithery/cli install @ahujasid/ableton-mcp --client claudebrew install uv{
"mcpServers": {
"AbletonMCP": {
"command": "uvx",
"args": ["ableton-mcp"]
}
}
}uvx ableton-mcpFAQs
How do I troubleshoot connection issues between the AbletonMCP server and the Remote Script on localhost:9877?
Check that Ableton's Control Surface is set to AbletonMCP in Preferences → MIDI. Verify the Remote Script loaded by checking Ableton's logs. Test port 9877 availability using netstat or lsof to confirm no other process is blocking it. Restart both Ableton and the MCP client after configuration changes, as the socket connection initializes when Ableton launches the Remote Script.
What specific Ableton Live API actions can AbletonMCP perform through natural language commands?
AbletonMCP translates natural language into Live API actions including creating tracks, writing MIDI clips with specific note patterns, loading devices from the browser, adjusting session tempo, controlling transport playback, applying effects like reverb, modifying note velocities for humanization, merging clips, and duplicating patterns across tracks. These operations execute through the Remote Script's access to Live's internal Python API.
Can I modify the AbletonMCP Remote Script to use a different TCP port than 9877?
Yes, modify the Remote Script's __init__.py file to change the hardcoded port from 9877 to your preferred port. You must also update the MCP server configuration to match. Ensure both components use identical port values. Non-standard ports may complicate troubleshooting since community support assumes the default 9877 configuration.
What are the latency and stability limitations of using AbletonMCP for real-time mixing during a live performance?
Real-time latency depends on your Python runtime overhead, local network stack performance, and Ableton's Remote Script thread priority. Stability is constrained by TCP's reconnection behavior, lack of error recovery in the single-threaded socket listener, and Python's garbage collection pauses. Test under load with concurrent parameter automation to measure actual jitter before deploying live.
Can MCP servers be integrated with other AI frameworks like TensorFlow or PyTorch?
Yes, MCP servers integrate with TensorFlow or PyTorch indirectly through agent frameworks like LangChain, LlamaIndex, or CrewAI. These frameworks use the models for inference or tool execution while MCP handles agent-to-tool communication. Build custom MCP servers to expose TensorFlow or PyTorch models as callable tools, or containerize model endpoints that MCP-compatible agents can discover and invoke at runtime.
How does MCP compare to traditional API gateways?
MCP is an AI-native bidirectional protocol enabling agents to dynamically discover tools via JSON-RPC, optimizing stateful interactions and reducing development effort through automatic capability querying. Traditional API gateways route stateless HTTP traffic with manual configuration, excelling in low-latency transactions but lacking native context retention or tool auto-discovery. Choose MCP for multi-agent ecosystems; use gateways for transactional applications.