Analysis updated 2026-05-18
Convert a photo of any object into a LEGO brick model with step-by-step building instructions
Type a description of what you want to build and get a 3D LEGO model to view and download
Download the LDR/MPD file from a generated model to open in BrickLink Studio or other LEGO design tools
Study how the pipeline connects image-to-3D AI, voxelization, and brick optimization as a full-stack example
| jjohnson5253/brickbuilderai | alsgur9865-sketch/second-brain-engine | compumaxx/gba-video-studio | |
|---|---|---|---|
| Stars | 10 | 10 | 10 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | moderate | hard |
| Complexity | 4/5 | 3/5 | 4/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires a fal.ai API key, high-quality SAM-3D mode needs a RunPod GPU endpoint (H100 recommended) with the published Docker image jjohnson5253/manifold-sam3d.
BrickBuilder is a web app that turns a photo or text description into a buildable LEGO-compatible brick model. You upload an image or type what you want, and within about 30 seconds you get a 3D brick model you can rotate in the browser, view step-by-step building instructions for, download as an LDR or MPD file that works with LEGO digital design tools, or use as a parts list to order physical bricks. The process goes through several steps. An AI model first reconstructs a 3D shape from the input. Two options are available: Trellis for simpler use, and SAM-3D for better quality (SAM-3D requires a separate GPU-powered worker hosted on RunPod). The 3D shape is then broken down into voxels (a 3D grid of cubes), and an optimizer converts those voxels into real LEGO-compatible bricks that fit together correctly. The project is split into three parts: a React and TypeScript frontend with a 3D viewer built on Three.js, a Python FastAPI backend that handles image processing and brick optimization, and an optional serverless worker for the SAM-3D pipeline running on a cloud GPU. Optional integrations include Supabase for storage and Stripe for payments, but both are optional: a local PostgreSQL database runs automatically if Supabase is not configured. To run it locally you need Python 3.10 or later, Node.js, the uv Python package manager, and a free API key from fal.ai. Two commands handle installation and startup. The higher-quality SAM-3D path requires deploying a public Docker container to RunPod with a GPU. The project is MIT-licensed and includes a note that LEGO is a trademark of the LEGO Group, which does not endorse this project.
A web app that converts any photo or text description into a buildable LEGO-compatible brick model, with a 3D viewer, building instructions, and downloadable files.
Mainly Python. The stack also includes Python, TypeScript, React.
Use freely for any purpose including commercial use as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Verify against the repo before relying on details.