TripoSG
An open-source AI 3D generator turns images or scribbles with text prompts into meshes locally on a CUDA GPU with at least 8GB VRAM.
TripoSG generates 3D meshes from visual references for people creating 3D assets or testing shape ideas. It accepts finished images as well as rough scribbles paired with a text prompt.
- Image-to-3D generation: Create shapes from photographs, cartoons or sketches, with geometry that follows the subject's appearance and structure.
- Scribble-based prototyping: Use TripoSG-scribble to turn a rough drawing and a written description into a 3D shape.
- Detailed geometry: Generate meshes with sharp edges, fine surface details and complex structures, including subjects with challenging topology.
- Mesh output controls: Export meshes as GLB files and limit their face count when you need simpler geometry.
The Python software runs locally and requires a CUDA-enabled GPU with at least 8GB of VRAM. It downloads generation weights and the RMBG-1.4 background-removal model, so obtaining those files requires an internet connection. An online demo is also available for the scribble model. The repository code is open source under the MIT license.