Repolode
Submit
lightningpixel

lightningpixel/modly

Desktop app that turns photos or text prompts into 3D meshes using open source AI models running on your local GPU.

View on GitHubWebsite
Stars
8,566
Forks
802
Open issues
106
Last commit
6 days ago

What is it?

Modly is a desktop application for generating 3D meshes from a single image, using open source AI models that run locally on your own GPU instead of a cloud service. It targets Windows, Linux, and Apple Silicon macOS.

Models are added through an extension system. Each extension is a GitHub repository with a manifest.json, and the app can install them directly from a URL. Official extensions cover Hunyuan3D 2 (Mini, Turbo, Fast), TripoSG, and Trellis2 GGUF. Workflows are built as node graphs, for example Image, then Generate Mesh, then Add to Scene, and imported meshes can be smoothed and decimated in the app.

The project also ships a stdlib-only Python CLI so agents and scripts can call a running desktop instance for health checks, model listing, workflow runs and mesh export. AMD GPUs are supported on Linux and Windows through ROCm.

Key features

  • 01Generates 3D meshes from images using local open source models on your GPU.
  • 02Installs model extensions directly from GitHub repository URLs.
  • 03Node-based workflow editor for chaining image-to-mesh steps into a scene.
  • 04Smooth and decimate imported meshes inside the app.
  • 05Python CLI for scripts and agents to drive a running Modly instance.
  • 06AMD GPU support via ROCm on Linux and Windows.

Use cases

  • Quickly turning reference photos into rough 3D meshes for prototyping or previs.
  • Generating 3D assets offline without sending images to a cloud service.
  • Automating image-to-mesh batches from scripts via the CLI.
  • Experimenting with different open source 3D generation models through extensions.

Quick start

  1. 1

    Clone the repository

  2. 2

    Install JS dependencies

    npm install
  3. 3

    Set up the Python backend environment and install requirements

    cd api && python -m venv .venv && pip install -r requirements.txt
  4. 4

    Run in development mode

    npm run dev

Things to know

  • Apple Silicon macOS is supported, not Intel Macs.
  • Generating meshes needs a capable local GPU, with AMD support via ROCm on Linux and Windows.
  • The license metadata is not clearly detected, though the README states MIT.

In collections

Self-hosted appsRun and train AI on GPUsDesktop GUI apps & toolkits

Topics

#3d#ai-local#image-to-3d#desktop-app#gpu#open-source#ai-tools#modly#self-hosted

FAQ

What is modly?
Desktop app that turns photos or text prompts into 3D meshes using open source AI models running on your local GPU. Modly is a desktop application for generating 3D meshes from a single image, using open source AI models that run locally on your own GPU instead of a cloud service. It targets Windows, Linux, and Apple Silicon macOS.
Is modly free to use? What license does it have?
modly is open source under the Other license. Check the license terms before using it in commercial products.
Is modly still maintained?
Actively maintained. The last commit to modly was on Nov 1, 2026.
What can I use modly for?
Quickly turning reference photos into rough 3D meshes for prototyping or previs. Generating 3D assets offline without sending images to a cloud service. Automating image-to-mesh batches from scripts via the CLI. Experimenting with different open source 3D generation models through extensions.
How do I install modly?
Clone the repository Install JS dependencies: npm install Set up the Python backend environment and install requirements: cd api && python -m venv .venv && pip install -r requirements.txt Run in development mode: npm run dev
Where can I get modly?
The source code and installation instructions are on GitHub: https://github.com/lightningpixel/modly

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects