What problem does it solve?
CAD-capable AI agents often generate build instructions as code but lack visual feedback, so they “guess” at geometry correctness and require slow human screenshot-and-retry loops.
Core Features & Use Cases
- Render immediate visual feedback: Send build123d modeling code to the server and get back rendered PNGs the agent can inspect.
- Create/modify models via API or MCP: Build, update, and enumerate models inside a containerized workflow suitable for automated iteration.
- Generate engineering-style views and exports: Request 3D renders, 2D technical drawings/blueprints, and export printable formats (STL/STEP/3MF), plus basic printability checks.
- Use Case: Iterate on a mechanical part (e.g., center a hole, adjust clearance, verify dimensions) by generating code, rendering, and refining based on the images and measurements—without manual screenshot relay.
Quick Start
Start the container server, then ask your AI to create a part by sending build123d code to POST /model/create and immediately request a visualization with POST /render/3d for the same model.