# livepilot [Health: Active]

**Category:** 🎥 Multimedia Process  
**Repository:** https://github.com/dreamrec/LivePilot  
**GitHub Stars:** 89  
**npm Downloads (last month):** 1781  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/livepilot

## Description
AI copilot for Ableton Live 12 — 91 MCP tools for music production and mixing

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "livepilot": {
    "command": "npx",
    "args": ["-y","livepilot"]
  }
}
```

## Documentation

## What the livepilot MCP server does

The livepilot MCP server gives AI clients an MCP interface to Ableton Live 12. It supports both precise editing and higher-level musical requests. Agents can control transport, tracks, clips, MIDI notes, devices, routing, mixing, automation, and Arrangement View. They can also search Ableton’s library, local plugins and racks, local samples, and Splice before loading material.

The tool surface covers 474 tools grouped into 57 domains. Beyond direct set control, the system includes audio inspection and creative assistance. It can read live spectral information, analyze rendered audio, compare a result with the state before an edit, preview alternatives, and help retain a preferred version. Project memory can track open creative threads, section roles, protected elements, and choices kept across sessions.

## How it works

LivePilot connects to Ableton through an installed Remote Script over TCP port 9878. A Max for Live bridge provides additional communication over UDP port 9880 and OSC port 9881. Local sample and Splice-related data uses SQLite for downloaded samples. The architecture also includes a device atlas, sample-processing knowledge, audio perception, corpus analysis, planning, evaluation, and memory systems.

Requests can be expressed as broad musical intent or exact parameters. For example, an agent may ask for a stylistic direction, or specify timing offsets, gain changes, frequency ranges, modulation, automation, or note behavior. Changes made inside the Live set use Ableton’s undo stack, allowing edits to be reversed through Live. Downloads, library scans, saved memory, and installer changes are outside that undo mechanism.

## Setup and configuration

The livepilot MCP server includes a setup wizard invoked with:

```bash
npx livepilot --setup
```

The wizard installs the Ableton Remote Script, creates the Python environment, copies the Max for Live Analyzer, and checks connectivity. Claude Desktop users can instead download the current `.mcpb` package from the project’s GitHub releases and open it directly.

For studio or live-performance systems, pinning a version is recommended because the project is changing quickly. Review the changelog before upgrading. The README identifies macOS and Windows as supported platforms.

## Tools and capabilities

- Control Live transport, tracks, clips, notes, devices, routing, mixing, automation, and arrangement.
- Search Ableton content, plugins, racks, local samples, and Splice before loading sounds.
- Inspect live spectral data and analyze rendered audio.
- Compare edits with earlier state and preview alternatives before keeping a result.
- Use device, drum-kit, sample-processing, and musical-technique knowledge.
- Preserve project context and learn from versions or moves retained across sessions.

_Full upstream README: https://allmcps.com/mcp/livepilot/readme_