DXN1-AUDIO-AI-TOOL
AI-powered music production toolkit enabling AI agents to generate, remix, separate stems, master tracks, and control DAWs via MCP.
README
<p align="center"> <img src="https://img.shields.io/badge/version-2.0.0-orange?style=for-the-badge" alt="v2.0.0"> <img src="https://img.shields.io/badge/license-MIT-green?style=for-the-badge" alt="MIT"> <img src="https://img.shields.io/badge/python-3.10+-blue?style=for-the-badge" alt="Python 3.10+"> <img src="https://img.shields.io/badge/tools-18_MCP_Tools-red?style=for-the-badge" alt="18 Tools"> </p>
DXN1-AUDIO-AI-TOOL
AI-powered music production toolkit. Generate music with Suno AI, remix in any DAW, separate stems, master, and more — all controllable by any AI agent (Claude, GPT, etc.) via MCP.
Install
# Clone the repo
git clone https://github.com/DXN1-termux/DXN1-DAW-AI-TOOL.git
cd DXN1-DAW-AI-TOOL
# Install dependencies
pip install -r requirements.txt
# (Optional) Install as a command-line tool
pip install -e .
That's it. dxn1 commands are now available.
Optional extras
# Stem separation (drums/bass/vocals/other) — needs GPU for speed
pip install demucs torch
# REAPER DAW control
pip install reapy
# Claude Vision (AI screenshot analysis of your DAW mixer)
pip install anthropic
# Suno AI music generation
pip install httpx
Set API keys (optional — local MIDI generation works without any keys)
export DXN1_SUNO_API_KEY="sk-..."
Quick Start
Generate music
# With Suno AI (needs API key)
dxn1 generate "Hardstyle banger with heavy distorted kicks at 150 BPM" --style hardstyle
# Without API key — generates a MIDI file locally
dxn1 generate "Dark trap beat" --local --bpm 140
# Generate MIDI only
dxn1 midi --root Am --scale NATURAL_MINOR --bpm 150 --bars 32 --genre hardstyle
Remix a song
# Analyze first
dxn1 analyze song.mp3
# Full remix (no DAW needed — pure audio processing)
dxn1 remix song.mp3 --style hardstyle --bpm 150
# Remix with DAW control (REAPER must be running)
dxn1 remix song.mp3 --style trap --backend reaper
# Separate stems first
dxn1 stem song.mp3
Master a track
dxn1 master track.wav --preset electronic
dxn1 master track.wav --preset hip_hop
Analyze audio
dxn1 analyze song.mp3
# → BPM: 128.0 Key: C minor LUFS: -12.3 Duration: 3:42
Use with AI (Claude Desktop, etc.)
Add to your Claude Desktop config:
{
"mcpServers": {
"dxn1-audio-ai-tool": {
"command": "python",
"args": ["-m", "dxn1_audio_ai_tool.server.mcp_server"]
}
}
}
Then just tell Claude: "Make me a hardstyle remix of this song at 150 BPM" — it knows exactly what to do.
All Commands
| Command | Description |
|---|---|
dxn1 generate "prompt" |
Generate music (Suno AI or local MIDI) |
dxn1 remix file.mp3 --style trap |
Full remix pipeline |
dxn1 analyze file.mp3 |
Detect BPM, key, loudness, duration |
dxn1 stem file.mp3 |
Split into drums / bass / vocals / other |
dxn1 master file.wav --preset modern |
Master a track (5 presets) |
dxn1 normalize file.wav --lufs -14 |
Normalize to target LUFS |
dxn1 midi --bpm 150 --genre hardstyle |
Generate a MIDI arrangement |
dxn1 serve |
Start MCP server for AI agents |
dxn1 dashboard |
Open real-time web dashboard |
MCP Tools (18 tools for AI agents)
AI Music Generation (Suno / Udio)
suno_generate— Generate a full song from textsuno_generate_custom— Custom lyrics with[Verse],[Chorus],[Drop]tagssuno_extend— Extend an existing generated tracksuno_remix— AI-powered remix via Sunogenerate_music— Auto-picks best provider (Suno → Udio → local MIDI)generate_and_import— Generate music AND load it into a DAW
Audio Processing
analyze_audio— BPM, key, loudness, durationseparate_stems— Drums / bass / vocals / other via Demucsnormalize_audio— LUFS normalizationpitch_shift— Shift pitch by semitonestime_stretch— Change tempo without changing pitchfade_audio— Add fade in / fade out
DAW Control
daw_connect— Connect to FL Studio / REAPER / Audacity / any DAWremix_song— Full remix pipeline in one call
Mixing & Mastering
apply_mix_preset— Genre-balanced mix (9 genres)list_mix_presets— List all available presetsmaster_audio— Full mastering chain (EQ → compression → limiting → LUFS)
Local Generation (no API key needed)
generate_midi— Complete MIDI song: chords, melody, bass, drums
Supported DAWs
| DAW | Cost | Controller | Best For |
|---|---|---|---|
| REAPER | Free trial / $60 | reapy Python API | All genres, all platforms |
| FL Studio | $99+ | pywinauto + MIDI Script | Electronic music, Windows |
| Audacity | Completely free | Keyboard automation | Basic editing, all platforms |
| Any DAW | — | Generic keyboard shortcuts | Quick tasks |
Environment Variables
| Variable | Default | Description |
|---|---|---|
DXN1_SUNO_API_KEY |
— | Suno AI API key |
DXN1_UDIO_API_KEY |
— | Udio API key |
DXN1_FL_EXE_PATH |
— | FL Studio executable path |
DXN1_OUTPUT_DIR |
./output |
Output directory |
DXN1_SERVER_TRANSPORT |
stdio |
MCP transport (stdio / sse) |
DXN1_VISION_ANTHROPIC_API_KEY |
— | Claude API for vision QC |
Project Structure
dxn1_audio_ai_tool/
├── audio/engine.py # Analysis, stem separation, normalize, pitch, stretch
├── services/
│ ├── suno_client.py # Suno AI — generate, extend, remix
│ ├── udio_client.py # Udio AI — generate, extend
│ └── ai_music_client.py # Unified client (auto-pick provider)
├── controllers/
│ ├── base.py # Abstract DAW interface
│ ├── fl_studio.py # FL Studio (pywinauto + MIDI Script)
│ ├── reaper_controller.py # REAPER (reapy)
│ ├── audacity_controller.py # Audacity (keyboard automation)
│ └── generic_daw.py # Any DAW (keyboard shortcuts)
├── pipeline/remix.py # Full remix orchestration
├── generation/midi_generator.py # Algorithmic MIDI songs
├── mixing/presets.py # 9 genre mix presets
├── mastering/chain.py # Offline mastering (EQ/comp/limit/LUFS)
├── vision/agent.py # Claude Vision screenshot analysis
├── feedback/
│ ├── dashboard.py # Real-time web dashboard
│ ├── overlay.py # Cursor tracking + screenshot annotation
│ └── action_log.py # Action history
├── server/
│ ├── mcp_server.py # 18 MCP tools for AI agents
│ └── system_prompt.py # System prompt that teaches AI how to use tools
├── cli.py # One command for everything
├── config.py # All settings + env var support
└── utils.py # Logging, timing, helpers
License
MIT
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。