Forge 3D

Live

AI-powered 3D model generation — prompt to printable object in minutes.

Source available on request

About this project

Overview

A desktop tool that takes text or image inputs, generates 3D models through Meshy AI and Tripo3D APIs, and prepares them for resin or FDM printing through an automated Blender cleanup pipeline.

It started as a personal tool for my 3D printing business. I was spending too much time jumping between browser tabs, downloading files manually, re-uploading them to Blender, fixing normals, setting origins, exporting STLs. The whole thing took 15 minutes of repetitive work per model. Forge3D collapses that into one window.

Technical decisions were shaped by real constraints. I used Electron with a single-file React app because I wanted a fast iteration cycle without a build server on my production machine. Three.js loads from CDN at runtime rather than bundled. The Blender integration runs as a separate git repo because the Python cleanup pipeline has nothing to do with the Node frontend.

Key features

What it does

Text and image to 3D generation via Meshy AI and Tripo3D, with Groq prompt enhancement

Automated headless Blender 5.0 cleanup — 3 quality levels, GLB and STL output

Batch generation queue for up to 5 prompts running sequentially

Multi-image to 3D using 2-4 reference angle photos for better geometry

Persistent generation history with canvas thumbnail capture and model reload

Prompt library with categories, saved settings, and keyboard shortcuts throughout

Built with

The Stack

React

Vite

Electron

Three.js

Groq

Meshy AI

What I'd do differently

In hindsight

I'd split App.jsx into components from day one. At 1500+ lines it works but adding features requires too much scrolling. I'd also set up a proper test suite earlier rather than relying on manual testing.

All projects

Want something like this built for you?

Tell me what you'd like to build and we'll see if it fits.

Start a Conversation