Summary
Automatic remesh and decimation for game- and web-ready assets, with materials preserved.
It leans on Instant Meshes for the retopology and pymeshlab for the reduction, wrapping both so an agent can analyze a model, act on the recommended target, and get back a GLB with its textures — the parts that usually mean bouncing between three separate desktop tools.
What it is
A server that hands 3D model processing to an assistant: it takes an OBJ or GLB file — local path, folder, or URL — and runs automatic retopology, face reduction, and mesh quality analysis, then outputs GLB.
What you get
- process_model reduces face count or rebuilds the mesh, with a mode that auto-picks between simplify and remesh, and options to preserve UVs and boundary features
- analyze_model reports face and vertex counts, topology problems like holes and separated components, and a recommended target face count
- manage_archives lists, cleans, copies, and inspects the archive folders each run produces — model, textures, and logs stored together
- Materials and texture maps are carried through the pipeline rather than dropped
Requirements
Windows only, since it drives the bundled Instant Meshes executable, and it needs a local Blender install for GLB conversion and texture extraction. Runs on Python.
Setup effort
Build from source — clone the repository and build it, then point your client at the binary