LiteModeler
Prompt to 3D. Local GPU. Production meshes.
AI-powered 3D model generation running entirely on your NVIDIA GPU. Type a prompt or drop in a reference image — LiteModeler generates production-quality GLB meshes with controllable polygon counts, texture resolution, and quality presets. Multi-model support (Trellis2 GGUF, Hunyuan3D), interactive Three.js viewer, and a built-in asset gallery.
Screenshot coming soon
What it does
Prompt-to-3D Pipeline
Type a text prompt — LiteModeler calls LiteImage to generate a reference image, then feeds it into the 3D reconstruction model. End-to-end prompt-to-GLB with zero manual steps.
Image-to-3D Generation
Drop in any reference image and generate a 3D mesh. Supports transparent backgrounds for cleaner reconstructions. The viewer shows real-time progress as the model builds geometry.
Multi-Model Support
Choose between Trellis2 GGUF (optimized for consumer GPUs), Hunyuan3D-2.1, and Hunyuan3D-Omni. Each model family has its own strengths — switch freely based on your quality/speed needs.
Granular Quality Control
Independent control over sparse steps (6–30), shape steps (6–30), texture steps (6–30), target face count (50K–500K polygons), and texture resolution (1024–4096px). Or use presets: Draft, Standard, High, Max.
Interactive 3D Viewer
Three.js-powered viewport with orbit controls, lighting, and real-time progress visualization. Inspect your generated models from any angle before exporting.
Asset Gallery
Every generated model is saved with metadata, thumbnails, and generation parameters. Browse with grid/list views, search, filter, and sort. Your entire 3D asset library in one place.
Setup Wizard
Guided installation flow — GPU detection, model family selection, Hugging Face authentication for gated models, and automatic weight downloads. Get generating in minutes.
Why LiteModeler
The only local desktop app with multi-model 3D generation, granular quality control (sparse/shape/texture steps), and integrated prompt-to-3D via LiteImage — not a cloud API, not a Blender plugin.
See it in action
Demo video coming soon
Requirements
- Windows 10/11 (64-bit)
- NVIDIA GPU with 12+ GB VRAM (CUDA required)
- 100+ GB free disk space for model weights
- Python 3.8+ (for inference runners)
Ready to try LiteModeler?
LiteModeler is available as a standalone Windows download.