- Home
- AI 3D Model Generators
- InstantMesh
InstantMesh
What it is
InstantMesh is a free, Apache License 2.0 framework that generates a 3D mesh from one image using the LRM/Instant3D architecture. Published by TencentARC, it runs on Linux through local or self-hosted deployment. A Gradio demo starts with `python app.py`, while `run.py` provides command-line mesh generation. The default export is an OBJ mesh with vertex colors; an alternate texture-map export takes longer. The project uses rembg to segment the foreground and can skip that step when an input already has an alpha mask. Four sparse-view reconstruction model variants and a customized Zero123++ UNet for white-background image generation are included. Docker deployment is documented. Recommended requirements are Python 3.10 or later, PyTorch 2.1.0 or later, and CUDA 12.1 or later; two GPUs can reduce demo memory use. Training code is included, and community implementations are noted for ComfyUI, Replicate, and Google Colab.
For and against
For
- Free Apache License 2.0 implementation
- Image-to-3D mesh generation
- Local Gradio and command-line workflows
- OBJ output with vertex colors
- Docker deployment documented
Against
- Linux and self-hosted setup required
- Needs compatible GPU hardware
- Requires Python, PyTorch, and CUDA
- Training dataset is not provided
The verdict
InstantMesh offers a local, open-source route from a single image to an OBJ mesh, with both demo and command-line workflows. Its hardware and dependency requirements make it most suitable for technically prepared users.
Who it suits
It suits developers and technical artists who want local or self-hosted image-to-3D generation. Users should be comfortable installing Python, PyTorch, CUDA, and compatible GPU hardware.
Plans and prices
InstantMesh
Apache-2.0 licensed implementation · local installation requires Python, PyTorch, CUDA, and a compatible GPU
Read 8 Oct 2026 · github.com
On the record
- Command-line use
- The `run.py` script generates 3D meshes from images via the command line.github.com · read 8 Oct 2026
- Dataset limitation
- The makers state that they cannot provide the training dataset because of its size.github.com · read 8 Oct 2026
- Deployment
- The README documents running the demo locally or with Docker.github.com · read 8 Oct 2026
- Hardware and dependencies
- The makers recommend Python 3.10 or later, PyTorch 2.1.0 or later, and CUDA 12.1 or later; the demo can use two GPUs to save memory.github.com · read 8 Oct 2026
- Image processing
- The project uses rembg to segment the foreground object and supports skipping that step when the input image already has an alpha mask.github.com · read 8 Oct 2026
- Integration
- The README acknowledges community implementations for ComfyUI, Replicate, and Google Colab demos.github.com · read 8 Oct 2026
- License
- The repository includes an Apache License 2.0.github.com · read 8 Oct 2026
- Local demo
- The repository provides a local Gradio demo that can be started with `python app.py`.github.com · read 8 Oct 2026
- Maker
- The repository is published by TencentARC; the pages opened do not state its headquarters or founding year.github.com · read 8 Oct 2026
- Mesh output
- The default export is an `.obj` mesh with vertex colors; an option exports a mesh with a texture map and takes longer.github.com · read 8 Oct 2026
- Model variants
- The project provides four sparse-view reconstruction model variants and a customized Zero123++ UNet for white-background image generation.github.com · read 8 Oct 2026
- Purpose
- InstantMesh is a feed-forward framework for generating a 3D mesh from a single image using the LRM/Instant3D architecture.github.com · read 8 Oct 2026
- Training
- The repository includes training code for sparse-view reconstruction models and Zero123++ fine-tuning.github.com · read 8 Oct 2026
