Open Generative AI
github.com/anil-matcha/open-generative-ai- Category
- AI Tools
- Rank
- No. 108Tools index
- Type
- APP
- Builder
- anil-matcha
- GitHub
- 28.1k stars
- Latest release
- v2.0.0
- Date
About
Complete AI image and video generation platform with multiple models, self-hostable application.
What it does
A browser and desktop studio for creating images, videos, lip sync clips, and cinematic shots. Users select models, enter prompts, adjust generation controls, and download results. Cloud jobs go to MuAPI and are polled until completion. Desktop users can also configure local image models or connect to a Wan2GP server.
Why it's ranked here
The project offers unusually broad media workflows behind one interface, with practical prompt, camera, history, and download tools. Its strongest distinction is the mix of MuAPI-backed generation and local model options. The value is tempered by external service dependence, uneven feature availability, and unsigned desktop installers.
What's good
Generation jobs survive navigation because studio pages remain mounted, while pending request identifiers persist in browser storage. Image controls cover aspect ratio, resolution, quality, seeds, reference strength, and model-specific inputs. The local catalog spans compact image models, SDXL, and Wan2GP video choices. English and Simplified Chinese interface text is included.
Tradeoffs
Most cloud generation requires a MuAPI access key, and the client sends requests directly to that service. Local inference is desktop-only, requires downloading multi-gigabyte weights, and some video models need a separately running Wan2GP server. Workflow and agent studios are marked web-only. The macOS app is not notarized, and the Windows installer is not code-signed.
How to use it well
Use it when creators or small teams want one visual workspace for prompt-driven images, video, lip sync, and camera-oriented scene setup. Start with hosted generation, then add local image models when offline control matters. Treat it as a generation front end, not a fully independent model-serving stack or a complete automation platform.
Technical notes+
package.json defines a private MIT-licensed JavaScript workspace using Next.js, React, Vite, Electron, and Electron Builder, with macOS, Windows, and Linux packaging targets. src/main.js implements a show-and-hide router that retains mounted studio nodes. src/lib/muapi.js submits authenticated MuAPI jobs and polls prediction results, with longer polling for video. src/lib/pendingJobs.js and src/lib/uploadHistory.js persist job and upload metadata in localStorage. src/lib/localModels.js catalogs bundled sd.cpp image models and Wan2GP-served image and video models. src/lib/uploadProxyTarget.js restricts upload targets to HTTPS, approved S3 patterns or configured hosts, blocks IP literals and local hosts, and rejects executable or active-content file types.
Observed
- License
- MIT
- Primary language
- JavaScript
- Application surfaces
- Hosted browser application, local web build, and Electron desktop application
- Desktop platforms
- macOS on Intel and Apple Silicon, Windows x64, and Linux x64
- Packaging
- npm workspace with Next.js, Vite, Electron, and Electron Builder scripts
- Generation interfaces
- Graphical studios for images, video, lip sync, cinema, workflows, agents, and developer tooling
- Local model support
- Bundled sd.cpp image inference plus configurable Wan2GP image and video providers
- Repository structure
- Workspace packages cover studio, workflow builder, agents, and design agent components
Read from README.md, package.json, src/main.js, src/counter.js, src/lib/i18n.js, src/lib/muapi.js, src/lib/models.js, src/lib/localModels.js, src/lib/pendingJobs.js, src/lib/promptUtils.js, src/components/Header.js, src/lib/uploadHistory.js, src/components/Sidebar.js, src/components/AuthModal.js, src/lib/uploadProxyTarget.js.
What it can do
Generate AI images from text prompts
Text descriptions/prompts → AI-generated images
Generate AI videos from text prompts
Text descriptions/prompts → AI-generated videos with audio
Convert images to videos
Images and text prompts → AI-generated videos
Create lip-synced videos
Video/image and audio/text → Lip-synced video content
Generate cinema-quality video content
Text prompts and creative parameters → High-quality cinematic videos
Self-host AI generation platform
Application files and server setup → Local AI generation server instance
Automate media generation pipelines
Terminal commands and prompts → Generated and processed media files
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.