AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
MCP server for VCV Rack 2 - composes modular synth patches by emitting structured JSON, no GUI automation.
Translation pending
Developer toolsPython · TypeScriptGitHub @sandraschiSelf-installOwner unconfirmed
Paste into Claude or ChatGPT for installation instructions suited to your environment.
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Installable projects can be verified without a deployment domain. Use GitHub sign-in for a personal repository, or the repository verification file above for an organization repository.
Ask about other verification options
Loading the full README. The saved excerpt is shown below.
MCP server for VCV Rack 2 built around one insight: .vcv patch files are plain JSON, so an LLM can compose modular synth patches by emitting structured data — no GUI automation, ever.
Status: v0.1.0 — Webapp and Core complete. Patch generation, catalog, validation, OSC bridge, webapp dashboard, Playwright E2E tests, and agentic workflow all working. Manual gates P2 (round-trip in Rack) and P3 (OSC e2e) still need a human with Rack open.
• Backend: 10916 (FastMCP + FastAPI)
• Frontend: 10917 (Vite React, active)
• OSC control: delegates to osc-mcp port 10767Read the full README ↗MCP server for VCV Rack 2 built around one insight: .vcv patch files are plain JSON, so an LLM can compose modular synth patches by emitting structured data — no GUI automation, ever.
Status: v0.1.0 — Webapp and Core complete. Patch generation, catalog, validation, OSC bridge, webapp dashboard, Playwright E2E tests, and agentic workflow all working. Manual gates P2 (round-trip in Rack) and P3 (OSC e2e) still need a human with Rack open.
git clone https://github.com/sandraschi/vcv-rack-mcp
cd vcv-rack-mcp
uv sync
uv run -m vcv_rack_mcp
| Tool | What it does |
|---|---|
vcv_patch | Create, edit, validate, list, get, open_in_rack, rack_cycle, import patches, set_module_data |
vcv_catalog | Search catalog, get module details, verify installed, library link, sideload, suggest rack |
vcv_live | OSC address map, performance sheet for live tweaking |
vcv_agentic_workflow | ctx.sample loop: brief → generate → validate → retry (max 3) |
show_patch_card / show_catalog_card | Prefab UI cards |
vcv-rack-mcp (port 10916, stdio / HTTP /mcp)
├── catalog/modules.yaml — 50 modules, 50/50 generative/performance
├── depot/ — .vcv patches + SQLite metadata + sidecar .md
└──→ osc-mcp (port 10767) — vcv_manager consumes address maps
| Feature | Status | Gate |
|---|---|---|
| Catalog (50 modules, incl. Stoermelder TRANSIT) | Done, validated | P1 |
| Patch generation | Done, deterministic | P2 (needs Rack round-trip) |
| Validation (4 checks, incl. module data) | Done with tests | P2 |
| OSC address map | Done | P3 (needs osc-mcp e2e) |
| Agentic workflow | Done | P4 |
| Webapp | Done, with Playwright E2E tests | P5 |
| Tauri wrapper | Future | P5 |