ai-mesh
A multi-provider CLI that turns a text prompt or a reference image into a textured 3D mesh.
ai-mesh is the mesh half of a composable pair with ai-img. Point it at an image, or hand it a prompt, and it submits the job to a hosted generator, waits, and writes a GLB to disk. Nothing runs on your GPU.
Providers are hosts and models are picked separately. On Fal it drives Tencent Hunyuan3D, Pixal3D, Tripo, and Rodin. On Meshy it handles both text-to-3D and image-to-3D. Target face count and PBR material maps are flags, and API keys can live in a config file instead of your shell.
It was built to be scripted, batched, and called by agents: one mesh per invocation, the output path on stdout, progress on stderr. Slow jobs can be submitted with --no-wait and collected later with ai-mesh fetch.