supercollider-mcp
Bridges MCP clients to a running SuperCollider IDE, enabling code evaluation, state queries, post window capture, and screenshots via OSC.
README
supercollider-mcp
MCP bridge to a running SuperCollider IDE. Lets an MCP client (Claude Code, Cursor, Continue, etc.) evaluate SC code, query values, capture the post window, and screenshot Qt windows without the user having to paste anything.
Sits alongside the other Ziforge MCPs for the EMS Stockholm rig — see Related MCPs.
How it works
sclang has a built-in OSC listener on port 57120. The setup.scd this repo ships with installs an OSCdef on /mcpEval that runs whatever code the MCP sends. Results come back on /mcpReply on port 57121, keyed by a request id.
Once you evaluate setup.scd in your IDE once per SC session, every MCP-side sc_eval call goes through OSC — no need to keep pasting.
Claude ↔ MCP server ↔ OSC :57120 → sclang → OSC :57121 → MCP → Claude
Install
git clone https://github.com/Ziforge/supercollider-mcp
cd supercollider-mcp
pip install -e .
Or with uv:
uv pip install -e .
Wire into Claude Code
{
"mcpServers": {
"supercollider": {
"command": "supercollider-mcp"
}
}
}
Setup — one line per SC session
Open your SC IDE, evaluate this once:
"~/Projects/supercollider-mcp/setup.scd".standardizePath.load;
You should see in the post window:
supercollider-mcp: listener installed on /mcpEval + /mcpPost (port 57120).
supercollider-mcp: replies sent to /mcpReply on 57121.
That's it. Now MCP calls work.
Tools
| Category | Tool | What it does |
|---|---|---|
| Eval | sc_eval(code, timeout) |
Send SC code, get result string |
| Eval | sc_query(expression, timeout) |
Same as sc_eval but semantic — for reading state |
| Eval | sc_post_recent(n) |
Last N eval results + errors from the OSC buffer |
| Patterns | sc_hush() |
Pdef.all.do(_.stop) |
| Server | sc_boot_server() |
Boot Server.default |
| Server | sc_quit_server() |
Quit Server.default |
| Interp | sc_recompile() |
Recompile class library (destroys the OSC listener — re-run setup.scd after) |
| Windows | sc_list_windows() |
Enumerate sclang / SuperCollider windows via osascript (macOS) |
| Windows | sc_screenshot(save_path) |
screencapture full screen (macOS) — the sclang Qt windows appear on it |
| Setup | sc_setup_listener_scd_path() |
Return absolute path to setup.scd for the current install |
Limitations
sc_recompilekills the listener. After a class-library recompile, the OSC listener is gone. You must re-evaluate setup.scd. There's no way around this — recompile reboots the interpreter.- Post window is partial.
sc_post_recentonly shows results of MCP-issued evaluations, not arbitrary.postlnoutput from patterns or synths. sclang doesn't expose its Post class's output stream over OSC. If you want to capture that too, wire your Pdefs to emit viaNetAddr("127.0.0.1", 57121).sendMsg("/mcpReply", 0, 1, "...")explicitly. - Screenshots are full-screen. Window-ID targeting on macOS needs assistive access permission most sessions don't have. If your terminal has been granted assistive access via System Settings → Privacy → Accessibility, you can tighten this later.
Example prompts (for the MCP client)
- "Evaluate
SinOsc.ar(440).playin SC and hush after 2 seconds" - "What's the current value of
~ws.destNamein SC?" - "Screenshot the SC IDE and describe what the bus view is showing"
- "Recompile the class library, then re-install the mcp listener"
Related MCPs
| Repo | For |
|---|---|
superclean-mcp |
SuperClean pattern control (Pdef.play, hush, list) — layered on top of this |
phase8-mcp |
Korg Phase 8 |
disting-nt-mcp |
Expert Sleepers Disting NT |
fh2-mcp |
Expert Sleepers FH-2 |
es9-mcp |
Expert Sleepers ES-9 |
erae-mcp |
Embodme ERAE Touch II |
elektron-mcp |
Elektron devices (fork of zerubeus) |
License
GPL v2. See LICENSE.
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.