# FL Studio AI — AI Beat Maker & Music Production Engine > **AI-powered beat generation, synthesis, mixing, and mastering — with MCP server integration for Claude Code and OpenCode** [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](./LICENSE) [![Stars](https://img.shields.io/github/stars/IzzoSol/FL-STUDIO-AI?style=social)](https://github.com/IzzoSol/FL-STUDIO-AI) [![Status](https://img.shields.io/badge/status-pre--release%20v8.0-blueviolet)](https://github.com/IzzoSol/FL-STUDIO-AI) [![Python](https://img.shields.io/badge/python-3.8%2B-blue)](https://python.org) --- ## What It Is **FL Studio AI** is a modular Python music production system that generates complete beats, melodies, chord progressions, basslines, and synthesizer patches — no manual piano roll entry required. It works three ways: - **MCP Server** — expose music tools (`generate_beat`, `synthesize_sound`, `mix_track`) directly to any MCP-compatible AI assistant (Claude Code, OpenCode). Ask your AI to make a beat; it calls the server and returns MIDI data ready for FL Studio. - **REST API** — Flask server on `localhost:5000`. HTTP endpoints for every feature, curl-friendly, integrates with any frontend or script. - **CLI / Python** — Direct command-line invocation for automation and batch workflows. - **Browser DAW** (`studio.html`) — Standalone web UI with beats, synth, drum machine, sequencer, mixer, and export — no backend required. > **Status:** Pre-Release v8.0 — 100+ Python modules, 20/20 core tests passing, all systems operational. --- ## Features (What Works Today) ### Beat & Track Generation - 15+ genre styles: trap, house, hip hop, techno, lo-fi, dubstep, drum & bass, UK drill, garage, jungle, ambient, and more - Natural language input: `python beatmaker.py make "dark trap beat in A minor 160 bpm"` - Full song structure (intro, verse, chorus, outro) via `auto_creator.py` - MIDI export ready for FL Studio piano roll ### Synthesis Engine - 47+ presets across 11 categories (leads, bass, pads, plucks, keys, strings, brass, FX, bells, SFX, special) - 5 synthesis methods: Wavetable, FM, Granular, Additive, Physical Modeling - Additional methods: Karplus-Strong, Modal, Vector, Analog Modeling, Circuit Modeling - ADSR envelopes, LFO modulation, filter bank, real-time arpeggiator - 47 virtual instruments across bass, keys, pads, leads, drums, FX ### Drum Machine - 5 kits: 808, trap, house, hip hop, techno - Polyrhythms, polymeter, swing, humanization, groove templates - 200+ patterns, 100+ styles in the pattern library ### AI Composition - `ai_melody_engine.py` — neural melody and chord generation with emotion profiles - `neural_music_generator.py` — LSTM-style composition with style transfer - `music_theory_engine.py` — 50+ scales, 100+ chord types, voice leading - `advanced_ai_composer.py` — 8 AI models: Markov Chain, LSTM, Style Transfer, Arrangement AI, and more ### Mixing & Mastering - 8-channel mixer with EQ, compression, sends/returns, metering - Stem separator — extract drums, bass, vocals, melody from audio - Vocal processor — pitch, harmonizer, vocoder - Mastering engine — 6 modes: analog, modern, vinyl, cassette, tape, digital. LUFS normalization - Effects rack — EQ, compressor, saturator, delay, reverb, chorus, flanger, phaser, distortion - Sidechain compression with envelope shaping ### Analysis Tools - BPM detection, key detection, pitch tracking - FFT spectrum analysis - Audio format converter ### Creative / Esoteric (Early / Experimental) - `esoteric_music_engine.py` — sacred geometry tuning, 528Hz Solfeggio frequencies, Solfeggio healing, binaural beat generation - `music_color_cymatics.py` — frequency-to-color mapping, cymatic pattern visualization ### Browser DAW (`studio.html`) - Single-file, zero-install — open in any browser - Tabs: Beats, Synth, Drums, Sequencer, Mixer, Export, Settings - Piano roll canvas, real-time audio via Web Audio API - LLM prompt-to-music (connects to Ollama / Groq / OpenAI / Anthropic / Gemini via settings) - MIDI / WAV export --- ## Quick Start ### 1. Start the API Server ```bash python flstudio_ai_api.py # Server runs at http://localhost:5000 ``` ### 2. Generate a Beat (CLI) ```bash # Style + tempo + bars python beatmaker.py generate trap 150 16 # Natural language python beatmaker.py make "dark lofi beat in A minor" # Full production (structure + mix + master) python auto_creator.py ``` ### 3. Use the Browser DAW (no server needed) ``` Open studio.html in your browser ``` ### 4. Start the MCP Server (for Claude Code / OpenCode) ```bash python flstudio_mcp_server_v2.py # or START_FLSTUDIO_MCP.bat ``` Then ask your AI assistant: ``` "Make a trap beat at 160 BPM in A minor" "Generate a chord progression in D minor" "Start playback in FL Studio" ``` --- ## MCP Tools When the MCP server is running at `localhost:5000`, these tools are available to any MCP client: | Tool | Description | |------|-------------| | `generate_track` | Full track — drums, bass, melody, chords | | `generate_drums` | Drum pattern in any style | | `generate_bass` | Bassline with scale and pattern | | `generate_melody` | Melody with scale and rhythm density | | `generate_chords` | Chord progression with voicing | | `flstudio_play` | Start FL Studio playback | | `flstudio_stop` | Stop playback | | `flstudio_record` | Toggle recording | | `flstudio_new_pattern` | Create new pattern | | `flstudio_piano_roll` | Open piano roll | | `get_note_data` | Retrieve all note data | | `get_midi_notes` | Retrieve MIDI format note data | **Example call:** ```json POST /mcp/call { "tool": "generate_track", "arguments": { "style": "trap", "tempo": 150, "key": 57, "scale": "minor", "bars": 16 } } ``` --- ## REST API Reference ```bash # Generate a full track curl -X POST http://localhost:5000/generate/track \ -H "Content-Type: application/json" \ -d '{"style":"house","key":60,"scale":"minor","bars":8,"tempo":128}' # Generate drums only curl -X POST http://localhost:5000/generate/drums \ -d '{"style":"trap","bars":2}' # Analyze key + BPM curl -X POST http://localhost:5000/analyze/key curl -X POST http://localhost:5000/analyze/bpm # Mastering suggestion curl http://localhost:5000/master/suggest ``` Key endpoints: `/generate/track` · `/generate/drums` · `/generate/bass` · `/generate/melody` · `/generate/chords` · `/generate/arps` · `/analyze/key` · `/analyze/bpm` · `/automation/volume` · `/samples/search` · `/creative/variations` · `/creative/fills` --- ## Tech Stack | Layer | Technology | |-------|-----------| | Language | Python 3.8+ | | API server | Flask | | Audio | numpy, scipy (DSP), soundfile | | MIDI | python-midi, mido | | MCP protocol | HTTP/JSON (Model Context Protocol) | | FL Studio bridge | OSC (python-osc), pyautogui | | Browser DAW | Vanilla HTML/CSS/JS, Web Audio API | | AI composition | Custom Markov Chain, LSTM-style, style transfer | | LLM integration | Ollama (local), Groq, OpenAI, Anthropic, Gemini (optional) | **Dependencies (install once):** ```bash pip install flask python-osc pyautogui numpy scipy soundfile mido ``` --- ## Verification & Tests ```bash # Quick test all 19 core modules python debug_verify.py # Comprehensive test suite (20/20) python final_test_100.py # End-to-end production test python e2e_test.py # Extended test suite python test_suite_v2.py ``` Individual module tests: ```bash python test_synth_quick.py python test_mixer_quick.py python test_analysis_quick.py ``` --- ## Project Structure ``` AI FL STUDIO BUILD/ ├── flstudio_ai_api.py Main REST API (Flask, port 5000) ├── flstudio_mcp_server_v2.py MCP protocol server (AI IDE integration) ├── beatmaker.py Simple CLI beat generator ├── auto_creator.py Full song generation ├── studio.html Browser DAW (single-file, no server needed) ├── launcher.py / launcher_web.html Launcher UI │ ├── super_engine.py Beat generation, 15+ styles ├── ai_melody_engine.py Neural melody composition ├── neural_music_generator.py LSTM-style + style transfer ├── music_theory_engine.py 50+ scales, 100+ chord types ├── advanced_synth.py 47+ presets, 5 synthesis methods ├── extensive_synth_engine.py 7 extended synthesis methods ├── drum_machine_v3.py 5 kits + patterns + fills ├── mixer_v3.py 8-channel mixer ├── mastering_engine.py 6 mastering modes + LUFS ├── effects_rack.py 9-effect chain ├── stem_separator.py Drum/bass/vocal/melody extraction ├── audio_analyzer.py BPM, key, FFT spectrum ├── esoteric_music_engine.py Sacred geometry, 528Hz, binaural (experimental) ├── music_color_cymatics.py Frequency-to-color visualization (experimental) │ ├── audio/ Audio assets and samples ├── exports/ Generated MIDI and audio outputs ├── presets/ Sound presets and configurations │ └── 00 INDEX.md – 22 MODULATION.md 22 reference documentation files ``` --- ## Roadmap - [ ] Direct FL Studio OSC write-back (patterns → FL Studio in real-time) - [ ] Persistent project save/load (`.flp`-adjacent format) - [ ] SHADDAI Orchestration Layer (SOL) — 7-agent AI council controlling the music engine - [ ] Pattern learning from user history (`23 PATTERN LEARNING.md`) - [ ] Mobile-optimized browser DAW - [ ] HuggingFace Spaces demo deployment - [ ] VST plugin wrapper (long-term) > Early/experimental modules are marked in the Features section above. The core generation, synthesis, mixing, and API pipeline is operational. --- ## License MIT — Use freely. Build legacy. ---

Built by IzzoSol  ·  @shaddaiAI  ·  Part of the ⚡ SHADDAI ecosystem