AI music generation from your coding agent
Create, iterate, and manage AI music directly from Claude Code, Cursor, or any MCP client.
AI music generation is powerful but disconnected from development workflows. This MCP server bridges the gap — generate music, manage tracks, and iterate on compositions without leaving your coding agent.
You: "Create an ambient electronic track for my app's loading screen"
→ Suno MCP generates track with optimized prompt
→ Returns audio URL + metadata
→ You integrate directly into your project
- Generate tracks with optimized Suno prompts from natural language
- Genre expertise built-in — 50+ genre prompt templates
- Iterate on existing tracks (extend, remix, style transfer)
- Manage your Suno library from the CLI
- Batch generate multiple variations
- Commercial licensing metadata tracking
npm install -g @frankxai/suno-mcp-serverOr add to your MCP config:
{
"suno-mcp": {
"command": "npx",
"args": ["@frankxai/suno-mcp-server"],
"env": {
"SUNO_API_KEY": "your-key"
}
}
}| Tool | Description |
|---|---|
suno_generate |
Generate a new track from description |
suno_extend |
Extend an existing track |
suno_remix |
Create variation of existing track |
suno_list |
List your generated tracks |
suno_get |
Get track details and download URL |
suno_prompt_optimize |
Optimize a natural language description into Suno prompt format |
The server includes built-in prompt optimization for Suno:
Input: "relaxing piano music for a meditation app"
Output: "[Ambient Piano] Gentle flowing arpeggios, soft sustain pedal,
warm reverb, 60 BPM, meditation-friendly, no vocals,
clean production, stereo width"
Genre templates cover:
- Electronic (ambient, synthwave, techno, house, DnB)
- Hip-hop (boom bap, trap, lo-fi, conscious)
- Rock (indie, post-rock, shoegaze, metal)
- Classical (orchestral, piano solo, chamber)
- World (afrobeat, latin, celtic, middle eastern)
- Experimental (glitch, IDM, noise, drone)
This MCP server integrates with the Agentic Creator OS music-lab plugin for full music production workflows.
Early development. Core MCP interface and prompt optimization engine are being built. Contributions welcome.
MIT