TripoSR

Free plan LinuxSelf-hostedWeb

What TripoSR is

TripoSR is an open-source model that reconstructs a 3D object from a single image. Tripo AI and Stability AI developed it collaboratively. The project reports generation in less than 0.5 seconds on an NVIDIA A100 GPU; default inference options use about 6 GB of VRAM for one image. The repository provides a command-line inference script and a local Gradio app, which can be started with `python gradio_app.py`. The script accepts one or more image paths. Its output can use a baked texture instead of vertex colors, with configurable texture resolution. The Gradio app includes background removal, foreground-ratio and marching-cubes-resolution controls, and OBJ and GLB outputs. Local setup calls for Python 3.8 or later, PyTorch, and dependencies installed from `requirements.txt`; CUDA is optional when available. The codebase, pretrained models, and interactive demo are released under the MIT license. TripoSR is listed as free. It supports image-to-3D reconstruction and AI texturing, but not text-to-3D. The project describes its intended audience as researchers, developers, and creatives working in 3D generative AI and 3D content creation.

Who it suits

TripoSR suits researchers, developers, and creatives working with 3D generative AI or 3D content creation. It is a fit for people who want to reconstruct 3D objects from images and can work with the listed local setup and hardware requirements.

What it does well

Reconstructs 3D objects from a single image.
MIT license covers the code, pretrained models, and interactive demo.
Includes a command-line inference script and local Gradio app.
Supports baked textures with configurable texture resolution.
Gradio app offers OBJ and GLB outputs.
CUDA is optional when available.

What to know first

Default inference uses about 6 GB of VRAM per image.
Local setup requires Python 3.8 or later, PyTorch, and listed dependencies.
Does not support text-to-3D.

Limits

The model and code are available under the MIT license and listed as free. Default inference options use about 6 GB of VRAM for one image input.

Plans

Plan	Billed and limits	Price
TripoSR	MIT-licensed model and code · default inference uses about 6 GB VRAM per image	Free

As github.com lists them, read 8 Oct 2026. The maker's own page has the final word.

Getting started

Open the project at https://github.com/VAST-AI-Research/TripoSR.
Install Python 3.8 or later, PyTorch, and the dependencies from `requirements.txt`.
Install CUDA if it is available; CUDA is optional.
Run inference with the command-line script using image path or paths.
For a local interface, start the Gradio app with `python gradio_app.py`.
Use the `--bake-texture` option to output a texture instead of vertex colors, and set texture resolution as needed.

Verdict

Choose TripoSR if you want an MIT-licensed model and local tools for image-to-3D reconstruction, with texture output and OBJ or GLB exports. Its stated speed applies to an NVIDIA A100 GPU, and default inference uses about 6 GB of VRAM per image. Look elsewhere if your workflow requires text-to-3D.

Questions

How much does TripoSR cost?

The listed TripoSR plan is free. The repository describes the model as open source under MIT.

Is TripoSR open source?

Yes. The codebase, pretrained models, and interactive demo are released under the MIT license.

What input does it support?

TripoSR reconstructs a 3D object from a single image. The inference script accepts one or more image paths.

Which output formats are available?

The Gradio app offers OBJ and GLB outputs. Inference can bake a texture instead of using vertex colors and allows texture resolution to be specified.

What do I need to run it locally?

The setup instructions list Python 3.8 or later, PyTorch, and dependencies installed from `requirements.txt`. CUDA is optional when available, and default inference options use about 6 GB of VRAM for one image.

Who developed TripoSR?

Tripo AI and Stability AI developed the model collaboratively. The MIT license file names both in its 2024 copyright notice.

Facts from TripoSR's own pages

Demo features: The Gradio app offers background removal, foreground-ratio and marching-cubes-resolution controls, and OBJ and GLB outputs. · github.com, 8 Oct 2026
Hardware requirement: Default inference options use about 6 GB of VRAM for a single image input. · github.com, 8 Oct 2026
How it works: The model generates a 3D model in less than 0.5 seconds on an NVIDIA A100 GPU, according to its README. · github.com, 8 Oct 2026
Input: The inference script accepts one or more image paths. · github.com, 8 Oct 2026
Intended users: The project describes its intended audience as researchers, developers, and creatives working in 3D generative AI and 3D content creation. · github.com, 8 Oct 2026
License: The codebase, pretrained models, and interactive demo are released under the MIT license. · github.com, 8 Oct 2026
License holders: The MIT license file names Tripo AI and Stability AI in its 2024 copyright notice. · github.com, 8 Oct 2026
Local installation: The README lists Python 3.8 or later, PyTorch, and dependencies installed from requirements.txt; CUDA is optional if available. · github.com, 8 Oct 2026
Local interface: The repository includes a Gradio app that can be started with python gradio_app.py. · github.com, 8 Oct 2026
Memory requirement: The default inference options use about 6 GB of VRAM for one image input. · github.com, 8 Oct 2026
Output options: The inference script can bake a texture instead of using vertex colors and lets users specify texture resolution. · github.com, 8 Oct 2026
Pricing: The TripoSR repository describes the model as open source under MIT, but does not state a subscription price for TripoSR. · github.com, 8 Oct 2026
Product identity: TripoSR is an open-source model for fast 3D reconstruction from a single image, developed collaboratively by Tripo AI and Stability AI. · github.com, 8 Oct 2026
Requirements: The setup instructions list Python 3.8 or newer and recommend installing CUDA if available. · github.com, 8 Oct 2026
Run locally: The repository provides a command-line inference script and a local Gradio app. · github.com, 8 Oct 2026
Single-image input: TripoSR reconstructs 3D objects from a single image. · github.com, 8 Oct 2026
Speed: The project says it can generate a 3D model in less than 0.5 seconds on an NVIDIA A100 GPU. · github.com, 8 Oct 2026
Texture output: Inference can output a texture instead of vertex colors using the --bake-texture option, with texture resolution configurable. · github.com, 8 Oct 2026
What it does: TripoSR reconstructs a 3D object from a single image using an open-source model. · github.com, 8 Oct 2026

Where TripoSR stands on Freedom251

AI 3D Model Generators#17 of 24