Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@MCP Panel Design a splitter silencer for a 24-inch duct"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
AI-powered CAD design copilot for acoustic products (splitter silencers and stratified acoustic panels). Built with FastAPI, Gemini, and MCP tools.
Docker and Docker Compose
Related MCP server: cadkit
1. Clone the repo
git clone <repo-url>
cd Task2. Create the .env file
cp .env.example .env # or create it manuallyAdd your key:
GEMINI_API_KEY=your_key_here3. Create the data directory
mkdir dataThis is where the SQLite database (data/projects.db) will be stored. It is mounted into the container so data persists across restarts.
4. Build and run
docker compose up --buildThe app will be available at http://localhost:8000.
python -m venv env
source env/bin/activate # Windows: env\Scripts\activate
pip install -r requirements.txt
uvicorn app:app --reloadapp.py FastAPI backend + Gemini agent loop
mcp_server.py MCP tool definitions
engine.py Silencer drawing engine
engine_panel.py Panel drawing engine
project_store.py SQLite-backed project persistence
email_intake.py Email intake router (parse, review, confirm/reject)
email_extract.py Gemini-powered parameter extractor
email_propose.py Proposal builder from extracted parameters
email_pending.py In-memory store for pending change proposals
data/ Mounted volume — contains projects.db
output/ Generated drawings (Docker named volume)
static/ Frontend assetsMethod | Path | Description |
GET |
| Health check |
POST |
| Send a message to the AI copilot |
GET |
| Retrieve a generated drawing |
Method | Path | Description |
POST |
| Receive an email ( |
GET |
| Browser UI listing all pending change proposals with confirm/reject buttons |
POST |
| Approve a pending proposal and commit the parameter change to the project |
POST |
| Discard a pending proposal without making any changes |
This server cannot be deployed
DXF and PDF/X-4 for AI agents: structured facts, PNG renders, an interactive in-chat viewer.
Parametric 3D CAD for AI agents: build print-ready parts, check them, export STL, 3MF or STEP.
AI Hub for AEC — 50+ 3D formats, clash detection, ACC integration via Autodesk Platform Services.
AI-callable calculators and engineering models with real formulas. No hallucinated math.
