Skip to content
SPECIFICTION
Get listed

InstantMesh

github.com

Free plan LinuxSelf-hosted
Visit github.com
Spec score 7.7 #17 of 24 in AI 3D Model Generators

What it is

InstantMesh is a free, Apache-licensed framework that generates a 3D mesh from one image using the LRM/Instant3D architecture. It runs on Linux in a local or self-hosted setup. A Gradio demo starts with `python app.py`, while `run.py` offers command-line generation. The default export is an OBJ mesh with vertex colors; an alternative adds a texture map and takes longer. rembg can segment the foreground object, or that step can be skipped when the source 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. The recommended environment uses 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, but the training dataset is unavailable because of its size. Community implementations are noted for ComfyUI, Replicate and Google Colab.

Spec sheet

Purpose
InstantMesh is a feed-forward framework for generating a 3D mesh from a single image using the LRM/Instant3D architecture.Read 8 Oct 2026 · github.com
Deployment
The README documents running the demo locally or with Docker.Read 8 Oct 2026 · github.com
License
The repository includes an Apache License 2.0.Read 8 Oct 2026 · github.com
Founded
2024Read 28 Sep 2026 · github.com
Maker
The repository is published by TencentARC; the pages opened do not state its headquarters or founding year.Read 8 Oct 2026 · github.com
Command-line use
The `run.py` script generates 3D meshes from images via the command line.Read 8 Oct 2026 · github.com
Dataset limitation
The makers state that they cannot provide the training dataset because of its size.Read 8 Oct 2026 · github.com
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.Read 8 Oct 2026 · github.com
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.Read 8 Oct 2026 · github.com
Integration
The README acknowledges community implementations for ComfyUI, Replicate, and Google Colab demos.Read 8 Oct 2026 · github.com
Local demo
The repository provides a local Gradio demo that can be started with `python app.py`.Read 8 Oct 2026 · github.com
Mesh output
The default export is an `.obj` mesh with vertex colors; an option exports a mesh with a texture map and takes longer.Read 8 Oct 2026 · github.com
Model variants
The project provides four sparse-view reconstruction model variants and a customized Zero123++ UNet for white-background image generation.Read 8 Oct 2026 · github.com
Training
The repository includes training code for sparse-view reconstruction models and Zero123++ fine-tuning.Read 8 Oct 2026 · github.com

Plans and prices

InstantMesh Free Apache-2.0 licensed implementation · local installation requires Python, PyTorch, CUDA, and a compatible GPU Read 8 Oct 2026 · github.com

Every InstantMesh plan and price →

Strong points and limits

Strong points

  • Generates meshes from a single image
  • Local Gradio and command-line workflows
  • OBJ export with vertex colors
  • Optional texture-map export
  • Apache License 2.0

Limits

  • Requires compatible GPU software stack
  • Training dataset is not provided
  • Texture-map export takes longer
  • Linux or self-hosted deployment

Who it suits

It suits developers and technical artists who can run Linux software with Python, PyTorch, CUDA and a compatible GPU. It is useful for image-to-3D workflows that need local or self-hosted execution.

Verdict

InstantMesh offers local image-to-3D generation with both graphical and command-line entry points. Its hardware and software requirements make environment preparation an important part of using it.

Where Specifiction places it

Alternatives to InstantMeshAll alternatives →