A toolset and docs let ChatGPT and Claude agents write LDraw source files that become LEGO CAD models, packaged as a dockerized web app.
A toolset and docs let ChatGPT and Claude agents write LDraw source files that become LEGO CAD models, packaged as a dockerized web app.
The agent takes a prompt, reads the repo's instruction docs, and plans the parts and submodels. It does not place bricks directly. It writes a plan file, then Python generator scripts that output the LDraw file. It uses tools for part search, collision and gap detection, and headless rendering, and loops render, inspect and adjust until it judges the model finished.
It produces LDraw source, several viewers, and a Blender-editable glb file. The author says only expensive high-end models can build large, correct models. Generation is slow, spaceships are weak, and VR on Meta Quest 3 has many issues.
The Docker build needs about 5 GB of disk. Without a TypeSafe API key, part search falls back to full-text search, which may give worse models.
Write an LDraw file (.ldr) for a LEGO model of [object, e.g. a small red sports car] using real LDraw part numbers and colors, and place the parts so they connect properly. Then validate the file, render a preview image if you can, and fix any floating or overlapping parts. Finish by giving me the final .ldr file and a parts list.
Score every demo request against your ICP, update HubSpot and ping the right rep in Slack within a minute of the form submit.
Every new Zendesk ticket gets a category, priority and sentiment tag plus a draft reply in an internal note before an agent opens it.
Copilot and Claude review each PR for bugs, missing tests and risky changes before a human looks, so reviewers can focus on design.