Developer Released AI Tool for LEGO CAD Designs

The LDraw Nova tool uses AI agents to automate the generation of CAD files for complex, virtual LEGO builds.

Updated on Oct. 4, 2026 in Artificial Intelligence

Isometric editorial illustration of a complex 3D structure built from interlocking plastic bricks floating in a clean, neutral space.
Developer Carlos Antelo released LDraw Nova, an open-source platform that uses AI agents to generate CAD files for complex virtual LEGO designs. AI Illustration. Upload story photo >

Live Poll

Do you trust AI agents to successfully design and build complex physical models?

Developer Carlos Antelo has released LDraw Nova, an open-source tool that employs AI agents to generate computer-aided design files for LEGO builds. The platform allows users to bypass manual brick placement by utilizing Python programs to create structural outputs.

Why it matters

The project represents a shift toward enabling AI agents to design complex, physical objects rather than purely digital content. By bridging AI models with CAD software, it provides a foundation for automated physical modeling.

The tool operates as a Docker web application supporting models from OpenAI, Anthropic, and OpenRouter, with costs estimated at $5 per Technic mechanism. While the agent manages iterative rendering to inspect and adjust designs, it currently lacks physics modeling.

The players

Carlos Antelo

The developer behind the LDraw Nova open-source project.

The details

LDraw Nova functions by having an AI agent draft a design plan in JSON, which is then translated into a Python program to produce LDraw source code. The agent iteratively renders the model as an image, allowing it to inspect and modify the design during the generation process. The system supports virtual reality viewing through the Meta Quest 3 headset, differentiating itself from the earlier, unrelated project known as LegoGPT.

Timeline

  1. October 2, 2026: Carlos Antelo officially released the LDraw Nova project.

The Tech Race

LDraw Nova shifts the focus from large-scale static training sets like the 47,000-structure BrickGPT repository toward iterative, agentic design workflows. It marks a departure from supervised learning models by prioritizing autonomous, multi-step construction planning.

The tool is available as an open-source Docker application on GitHub for users looking to automate LEGO CAD designs. While digital rendering and VR viewing on the Meta Quest 3 are currently possible, users must note that physical stability is not guaranteed.

The takeaway

The platform demonstrates a new method for AI-assisted physical prototyping through iterative code generation rather than traditional brick-by-brick modeling. Interested users should monitor the developer’s progress toward 3D-printing a test model to confirm the tool's real-world structural capability.

Further reading

For broader context on how agentic systems are changing creative workflows, visit Artificial Intelligence.

Source note: This article includes information reported by Tom's Hardware.

Live Poll

Do you trust AI agents to successfully design and build complex physical models?