Our take
Activeregular commits- Activity
- 40
- Community
- 41
- Issues
- 21
- Pull requests
- 21
This is an actively changing, multi-workflow creative interface, but adopting it means relying on Muapi.ai for generation and accepting a project with no recent release or test suite.
Good fit if
- You want a customizable image, video, cinema, and lip-sync interface and are comfortable using Muapi.ai's API.
- You prefer a hosted browser option or want to run the app with Node.js and build macOS or Windows desktop versions.
- You can work with an MIT-licensed project whose recent development is active and involves several contributors, without relying on a single dominant contributor.
Look elsewhere if
- You need generation to run entirely offline or want to avoid sending inputs to an external API service.
- Your workflow depends on a stable, versioned release or a repository with automated tests and CI.
- You need another generation provider, since this app's API integration is centered on Muapi.ai.
All health signals
| Last commit | 2026-11-07 (2 days ago) |
|---|---|
| Commits, last 90 days | 43 |
| Releases, last 12 months | 0 |
| Contributors | 4 (top contributor: 58% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 7 months |
Checked on 2026-11-09 with the GitHub API.
Overview
Open-Higgsfield-AI is an image, video, cinema, and lip-sync studio built on Muapi.ai, the API service that handles model requests. It brings Higgsfield-style creative workflows into a customizable app, with text-to-image, image-to-image, text-to-video, image-to-video, and audio-driven lip sync. The project describes itself as an alternative to a closed, subscription-based creative platform.
You can use the hosted browser version, run the app yourself with Node.js, or build desktop apps for macOS and Windows. Self-hosting gives you control over the interface and its local history, but generation still relies on Muapi.ai and requires its API key; uploaded media is sent to that service for processing.
Key Features
- Image and video studios switch between text-driven and reference-image workflows when you add an input image.
- Lip Sync Studio supports portrait-image-plus-audio and video-plus-audio workflows.
- Cinema Studio provides camera, lens, focal-length, and aperture controls that shape prompts.
- Compatible image models can accept multiple reference images, with selection and ordering controls.
- Generation history and uploaded-image history persist in browser storage.
- The shared
packages/studiolibrary serves both the self-hosted app and Muapi.ai's hosted version. - Electron build scripts produce desktop installers for macOS and Windows.
Use Cases
- An independent creator who wants one interface for experimenting with image and video models through a Muapi.ai account.
- A filmmaker or storyboard artist who wants to guide cinematic generations with camera-style controls.
- A creator making talking portraits or dubbing existing footage with audio-driven lip sync.
- A developer who wants to customize a React-based creative studio or reuse its shared studio components.
What you need
Detected in the repository
- Node.js (from package.json)
Getting Started
Install Node.js 18 or later and obtain a Muapi.ai API key, then run:
git clone https://github.com/Autom8AI/Open-Higgsfield-AI.git
cd Open-Higgsfield-AI
npm install
npm run dev
Open http://localhost:3000 and enter your API key. See the README for production and desktop builds, setup details, and the hosted option.
License in Plain Words
No standard license was detected. Without a license, the default copyright rules apply: you may read the code, but not reuse it. Check the repository or ask the authors.
Alternatives
- artcraft: ArtCraft is a Rust desktop app with compositing and 3D scene tools, while Open-Higgsfield-AI offers browser and desktop workflows around a broad model catalog.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| Open-Higgsfield-AI | JavaScript | – | 2.5k | Active |
| artcraft | Rust | Apache-2.0 | 32.1k | Active |
Considerations
- The repository has recent commits and is not archived, but it has no releases in the past year. There are no tests or CI configured in the reported project facts, and no issues closed or pull requests merged in the recent measurement window.
- Self-hosting does not make model inference local. Requests use Muapi.ai, and uploaded media is sent to its upload API before being passed to compatible models. Treat the API key and media handling accordingly.
- The README says the API key is kept in browser
localStorage; this is a BYOK arrangement, not a standalone model backend. - Desktop installers are described as unsigned or not notarized, so macOS Gatekeeper and Windows SmartScreen may show warnings during installation.
Found this useful?
Share it with someone who would like Open-Higgsfield-AI.