Skip to content

Audiomancer Documentation

MCP server for AI-assisted music production. Manages sample libraries, analyzes audio, generates patterns, and integrates with TidalCycles live coding.

What is Audiomancer?

Audiomancer gives AI assistants deep integration with music production workflows. Instead of just chatting about music, the AI can:

  • Browse and enable sample packs from your sample library without you leaving the conversation
  • Generate TidalCycles patterns using your actual sample library (not generic "bd" and "sn")
  • Find similar samples using audio embeddings and semantic search
  • Analyze audio files for BPM, key, spectral features, and more

It's designed for TidalCycles live coding but the analysis and search features work for any DAW workflow.

Architecture

Audiomancer Architecture Audiomancer Architecture

Quick overview:

  • Sample Library (source directory) contains your sample packs
  • Library Manager copies files to local cache and creates symlinks for SuperDirt
  • Audio Analyzers extract features and generate ML embeddings
  • Storage persists metadata (SQLite) and vectors (LanceDB)
  • Pattern Generators query samples and create TidalCycles code
  • MCP Server exposes 15 tools to AI clients

See Technical Architecture for detailed documentation.