FM8 MCP Server
Enables AI assistants to control Native Instruments FM8 synthesizer parameters through MIDI CC messages, allowing natural language programming of FM synthesis sounds including operator routing, frequency ratios, and modulation matrix settings.
README
FM8 MCP Server
AI-controlled Native Instruments FM8 synthesizer via MIDI and Model Context Protocol.
What This Does
Lets Claude Desktop (or other MCP clients) control your FM8 synthesizer by sending MIDI CC messages through a virtual MIDI port.
Requirements
- Node.js 18+
- Native Instruments FM8
- Virtual MIDI port software:
- Windows: loopMIDI
- macOS: IAC Driver (built-in)
Quick Start
git clone https://github.com/thesigma1receptor/fm8MCP.git cd fm8MCP npm install npm run build
Setup Virtual MIDI
Windows
- Install and run loopMIDI https://www.tobias-erichsen.de/software/loopmidi.html
- Remove any existing ports
- Create port named
fm8
macOS
- Open Audio MIDI Setup
- Window → Show MIDI Studio
- Double-click IAC Driver
- Enable "Device is online"
- Create port named
fm8
Configure FM8
- Set MIDI input to
fm8port by clicking off/on, change to on - Go to Master tab and load included MCP.f8c file
- Right click any operators that are off to enable them
Claude Desktop Setup
Edit your Claude config file:
Windows: %APPDATA%\Claude\claude_desktop_config.json
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Included in repo, or amend yours with:
{ "mcpServers": { "fm8": { "command": "node", "args": [ "C:/path/to/fm8-mcp-server/dist/server.js", "--transport=stdio", "--midi-port-name=fm8", "--channel=1" ] } } }
Restart Claude Desktop and you're ready!
AnythingLLM Setup
Requirements: Ollama with llama3.1 or qwen2.5:4b minimum
Edit: %APPDATA%\anythingllm-desktop\storage\plugins\anythingllm_mcp_servers.json
Included in repo, or amend yours with:
{ "fm8": { "name": "FM8 Control", "command": "node C:/path/to/fm8-mcp-server/dist/server.js", "args": ["--transport=stdio", "--midi-port-name=fm8", "--channel=1"], "env": {}, "cwd": "C:/path/to/fm8-mcp-server" } }
- Restart AnythingLLM
- Enable MCP server in workspace settings
- Use
@agentin chat to access tools
What You Can Control
FM8 Matrix (CC 0-62)
All operator-to-operator modulation routes (A→B, B→C, etc.)
Operator Outputs (CC 63-70)
Operator-to-Output levels (A→Out, B→Out, etc.)
Frequency Ratios (CC 72-77)
Harmonic/inharmonic relationships for each operator
See mappings.json for complete CC reference.
Example Claude Prompts
- "Create an FM bell sound"
- "Set operator A to modulate B at 75%"
- "Make a pure sine wave"
- "Route operator C to output at full volume"
- "Show all available FM8 controls"
If you want to give Claude free reign to try things:
-"Ratios are from 0 - 64:1, 127 being 64, feel free to look up fm synthesis techniques. There 8 operators, A-F are pure sine waves, X is a noise oscillator and Z is a filter. Let's see what you can come up with!"
Make it your own!
If you want to add any functionality, ie, more midi cc, use the standalone fm8, turn on midi learn and I used touchosc to make ~80 controls and map things. Look at mappings.json and add start with cc #78.
There are also ways with mcp to give claude memory and better instructions. So he could potentially learn how to make complex sounds.
Tip
This also works from VSCode with copilot.
Start server: node C:\fm8MCP\dist\server.js --transport=stdio --midi-port-name=fm8 --channel=1
Example command for agent:
node -e "const easymidi = require('easymidi'); const out = new easymidi.Output('fm8'); console.log('⚡ Creating neuro bass...'); out.send('cc', {controller: 0, value: 95, channel: 0}); setTimeout(() => out.send('cc', {controller: 1, value: 110, channel: 0}), 20); setTimeout(() => out.send('cc', {controller: 8, value: 105, channel: 0}), 40); setTimeout(() => out.send('cc', {controller: 9, value: 90, channel: 0}), 60); setTimeout(() => out.send('cc', {controller: 17, value: 100, channel: 0}), 80); setTimeout(() => out.send('cc', {controller: 63, value: 127, channel: 0}), 100); setTimeout(() => out.send('cc', {controller: 64, value: 110, channel: 0}), 120); setTimeout(() => { console.log('🎵 Neuro bass loaded! Play low notes (C1-C2)'); process.exit(0); }, 400);"
Troubleshooting
No MIDI signal
- loopMIDI/IAC running?
- FM8 input set to
fm8port? - MIDI channel matches (default: 1)?
Server won't start
- Run
npm run buildfirst - Check Node version:
node --version
Claude can't find it
- Use absolute paths in config
- Forward slashes even on Windows
- Restart Claude after config changes
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.