Tulasi — AI Native 3D Designer

Transforms a single photo into an editable, dimension-aware 3D model using generative AI, computer vision, parametric modeling, and browser-based rendering.

Turning a real-world object into an editable 3D model normally means manual CAD work or expensive scanning hardware — putting real 3D design out of reach for anyone without that budget or skill.

A working solo-founder pipeline that takes a single photo and produces an editable, dimension-aware 3D model, rendered and manipulable directly in the browser — no CAD software or scanning hardware required.

Solo founder — designed and built the full pipeline end-to-end: computer vision, generative AI integration, parametric modeling, backend, and the browser-based rendering UI.

Features

• Single-photo to 3D model conversion using generative AI and computer vision
• Dimension-aware reconstruction — output models carry real-world scale, not just shape
• Interactive browser-based model manipulation with real-time WebGL rendering
• Parametric editing so generated models can be adjusted, not just viewed

Built with OpenCV, Three.js, CadQuery, OpenCascade, MediaPipe, Python, Node.js, Supabase, Vercel, LLM APIs.

Source on GitHub