[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-11559":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":10,"archived":22,"fork":22,"defaultBranch":23,"hasWiki":24,"hasPages":22,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":43,"readmeContent":44,"aiSummary":45,"trendingCount":16,"starSnapshotCount":16,"syncStatus":46,"lastSyncTime":47,"discoverSource":48},11559,"Open-Generative-AI","Anil-matcha\u002FOpen-Generative-AI","Anil-matcha","Open-source alternative to AI video platforms — Free AI image & video generation studio with 200+ models (Flux, Midjourney, Kling, Sora, Veo). No content filters. Self-hosted, MIT licensed.","https:\u002F\u002Fdev.muapi.ai\u002Fopen-generative-ai",null,"JavaScript",18839,3201,144,6,0,99,774,5985,528,45,false,"main",true,[26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42],"ai-art-generator","ai-image-generation","ai-video-generation","creative-tools","flux-1","generative-ai","image-to-video","javascript","kling-ai","midjourney-alternative","muapi","open-source","sora-alternative","text-to-video","uncensored","unrestricted","wan-video","2026-06-12 02:02:32","# Open Generative AI — Open-Source Alternative to AI Video Platforms\n\n> **The free, open-source alternative to AI Video Platforms.** Generate AI images and videos using 200+ state-of-the-art models — no content filters, no closed ecosystem, no subscription fees.\n\n**Community:** Join [Reddit](reddit.com\u002Fr\u002Fmuapi) & [Discord](https:\u002F\u002Fdiscord.gg\u002FQhTrNRU4r3) for discussions and support\n\n> 🤖 **Automate media generations with AI coding agents:** [Generative-Media-Skills](https:\u002F\u002Fgithub.com\u002FSamurAIGPT\u002FGenerative-Media-Skills) — a library of skills that let agents like **Claude Code**, **Codex**, and other coding assistants drive 200+ image\u002Fvideo models end-to-end (prompt → generate → edit → stitch) directly from your terminal. Perfect for building automated media pipelines without touching a UI.\n\n### Related projects\n\n> **Open-source Node based workflow builder** -> https:\u002F\u002Fgithub.com\u002FSamurAIGPT\u002FVibe-Workflow\n\n> **Open-source AI Clipping — turn any long-form YouTube video into viral-ready vertical shorts** -> https:\u002F\u002Fgithub.com\u002FSamurAIGPT\u002FAI-Youtube-Shorts-Generator\n\n> **Open-source AI Design Agent** -> https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-AI-Design-Agent \n\n## 🌐 Try it Online — No Install Required\n\n**Hosted version:** [https:\u002F\u002Fmuapi.ai\u002Fopen-generative-ai](https:\u002F\u002Fmuapi.ai\u002Fopen-generative-ai)\n\nUse all four studios (Image, Video, Lip Sync, Cinema) directly in your browser — no Node.js, no setup. Sign up for a free account to start generating. The hosted version is always up to date with the latest models.\n\n**Follow** the [creator](https:\u002F\u002Fx.com\u002Fmatchaman11) for updates\n\n---\n\n## ⬇️ Download Desktop App\n\nOne-click installers — no Node.js or terminal required.\n\n| Platform | Download |\n|---|---|\n| macOS Apple Silicon (M1\u002FM2\u002FM3\u002FM4) | [Open Generative AI-1.0.9-arm64.dmg](https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-Generative-AI\u002Freleases\u002Fdownload\u002Fv1.0.9\u002FOpen.Generative.AI-1.0.9-arm64.dmg) |\n| macOS Intel (x64) | [Open Generative AI-1.0.9.dmg](https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-Generative-AI\u002Freleases\u002Fdownload\u002Fv1.0.9\u002FOpen.Generative.AI-1.0.9.dmg) |\n| Windows (x64) | [Open Generative AI Setup 1.0.9.exe](https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-Generative-AI\u002Freleases\u002Fdownload\u002Fv1.0.9\u002FOpen.Generative.AI.Setup.1.0.9.exe) |\n| Linux (Ubuntu x64) | [v1.0.9 release](https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-Generative-AI\u002Freleases\u002Ftag\u002Fv1.0.9) (`.AppImage` \u002F `.deb`), or build locally with `npm run electron:build:linux`. |\n\nAll releases: [github.com\u002FAnil-matcha\u002FOpen-Generative-AI\u002Freleases](https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-Generative-AI\u002Freleases)\n\n### macOS Installation Guide\n\nBecause the app is not notarized by Apple, macOS Gatekeeper will block it on first launch. Follow these steps:\n\n**Step 1** — Mount the DMG and drag the app to `\u002FApplications`\n\n**Step 2** — Open Terminal and run:\n```bash\nxattr -cr \"\u002FApplications\u002FOpen Generative AI.app\"\n```\n\n**Step 3** — Right-click the app in `\u002FApplications` → click **Open** → click **Open** again on the dialog\n\n> You only need to do this once. After that, the app opens normally.\n\n**Alternative (no Terminal):**\n1. Try to open the app — macOS will block it\n2. Go to **System Settings → Privacy & Security**\n3. Scroll down to find _\"Open Generative AI was blocked\"_\n4. Click **Open Anyway** → **Open**\n\n### Windows Installation — SmartScreen warning fix\n\nWindows SmartScreen may show a warning because the installer is not code-signed:\n\n1. Click **More info** on the SmartScreen dialog\n2. Click **Run anyway**\n\nThe app will install silently to `%LocalAppData%` with a Start Menu shortcut.\n\n### Ubuntu \u002F Linux Installation\n\nLinux artifacts are available when building with Electron Builder:\n\n```bash\n# Build Linux installers (AppImage + .deb)\nnpm run electron:build:linux\n```\n\nGenerated files are written to the `release\u002F` folder:\n- **AppImage** — portable, run directly after making executable:\n  ```bash\n  chmod +x \"release\u002FOpen Generative AI-*.AppImage\"\n  .\u002Frelease\u002FOpen\\ Generative\\ AI-*.AppImage\n  ```\n- **.deb** — install on Debian\u002FUbuntu:\n  ```bash\n  sudo apt install .\u002Frelease\u002Fopen-generative-ai_*_amd64.deb\n  ```\n\nIf AppImage fails to start on older systems, install `libfuse2`:\n\n```bash\nsudo apt install libfuse2\n```\n\n#### Ubuntu 24.04+ \u002F AppArmor sandbox restriction\n\nUbuntu 24.04 and later enable a kernel security policy (`apparmor_restrict_unprivileged_userns`) that blocks Chromium's user-namespace sandbox. If the app fails to start silently or crashes immediately, you have two options:\n\n**Option A — Recommended: install the `.deb` instead.**\nThe `.deb` package ships an AppArmor profile that grants the required permission automatically on install with no system-wide changes.\n\n**Option B — Temporary system fix (AppImage users):**\n```bash\nsudo sysctl -w kernel.apparmor_restrict_unprivileged_userns=0\n```\nThis lasts until next reboot. To make it permanent:\n```bash\necho 'kernel.apparmor_restrict_unprivileged_userns=0' | sudo tee \u002Fetc\u002Fsysctl.d\u002F99-userns.conf\n```\n\n---\n\nOpen Generative AI is a free, open-source AI image, video, cinema, and lip sync studio that brings creative workflows to everyone. No content filters, no prompt rejections, no guardrails — just full creative freedom. Powered by [Muapi.ai](https:\u002F\u002Fmuapi.ai), it supports text-to-image, image-to-image, text-to-video, image-to-video, and audio-driven lip sync generation across models like Flux, Nano Banana, Midjourney, Kling, Sora, Veo, Seedream, Infinite Talk, LTX Lipsync, Wan 2.2, and more — all from a sleek, modern interface you can self-host and customize.\n\n**Why Open Generative AI instead of other AI Video Platforms?**\n- **No filters** — no content filters, no nanny guardrails, no prompt rejections\n- **Free & open-source** — no subscription, no vendor lock-in\n- **Self-hosted** — your data stays on your machine, full creative control\n- **200+ models** — text-to-image, image-to-image, text-to-video, image-to-video, lip sync\n- **Multi-image input** — feed up to 14 reference images into compatible models\n- **Lip Sync Studio** — animate portraits or sync lips to any audio with 9 dedicated models\n- **Extensible** — add your own models, modify the UI, build on top of it\n\nFor a deep dive into the technical architecture and the philosophy behind the \"Infinite Budget\" cinema workflow, see our [comprehensive guide and roadmap](https:\u002F\u002Fmedium.com\u002F@anilmatcha\u002F).\n\n## ⚡ Local Model Inference (Desktop App Only)\n\nThe desktop app supports **two independent local engines**. Pick whichever fits the machine you actually run on:\n\n| Engine | What it is | Best for |\n|---|---|---|\n| **sd.cpp** (bundled) | C++ engine from [stable-diffusion.cpp](https:\u002F\u002Fgithub.com\u002Fleejet\u002Fstable-diffusion.cpp), runs on the same machine as the app. Metal GPU on Apple Silicon, CUDA\u002FVulkan\u002FROCm on Linux\u002FWindows. | Image-only models. Works on Mac M-series. |\n| **Wan2GP** (BYO server) | HTTP client to a user-run [Wan2GP](https:\u002F\u002Fgithub.com\u002Fdeepbeepmeep\u002FWan2GP) server. The server runs Python + PyTorch on a CUDA\u002FROCm GPU; the desktop app only sends prompts and receives results. | Video models (Wan 2.2, Hunyuan, LTX) and large image models (Flux, Qwen-Image). NVIDIA\u002FAMD GPU required on the *server*; the desktop app itself can run on a Mac. |\n\nBoth engines share the same UI: open **Settings → Local Models** to configure each.\n\n### Engine 1 — sd.cpp (bundled)\n\n| Model | Type | Size | Notes |\n|---|---|---|---|\n| **Z-Image Turbo** ⚡ | Diffusion Transformer | 2.5 GB + 2.7 GB aux | 8-step turbo. Heavy on memory. |\n| **Z-Image Base** ⚡ | Diffusion Transformer | 3.5 GB + 2.7 GB aux | 50-step high-quality. Heavy on memory. |\n| **Dreamshaper 8** | SD 1.5 | 2.1 GB | 20-step versatile. Lightest tested option on Mac. |\n| **Realistic Vision v5.1** | SD 1.5 | 2.1 GB | 25-step photorealistic |\n| **Anything v5** | SD 1.5 | 2.1 GB | 20-step anime\u002Fillustration |\n| **SDXL Base 1.0** | SDXL | 6.9 GB | 30-step high-res |\n\n> **Z-Image models** require two shared auxiliary files (downloaded once, shared across both models):\n> - **Qwen3-4B Text Encoder** — 2.4 GB\n> - **FLUX VAE** — 335 MB\n\n**How to use:**\n1. Open **Settings → Local Models** in the desktop app\n2. Install the **sd.cpp inference engine** (one click — auto-downloaded)\n3. Download your chosen model (and auxiliary files for Z-Image)\n4. In **Image Studio**, click the **⚡ Local** toggle next to the model selector\n5. Select your local model and generate — no API key needed\n\nAll downloads happen inside the app. Nothing is installed system-wide.\n\n### Engine 2 — Wan2GP (remote Gradio server)\n\nThe app does **not** bundle Python or model weights for Wan2GP. You run Wan2GP yourself on a machine with a CUDA or ROCm GPU and point the desktop app at its URL.\n\n```bash\n# On your GPU machine\ngit clone https:\u002F\u002Fgithub.com\u002Fdeepbeepmeep\u002FWan2GP\ncd Wan2GP\n.\u002Finstall.sh                          # or install.bat on Windows\npython wgp.py --listen --server-name 0.0.0.0   # binds to all interfaces\n```\n\nThen in the desktop app: **Settings → Local Models → Wan2GP server**, paste the URL (e.g. `http:\u002F\u002F192.168.1.42:7860`), click **Test**, then **Save**. The Wan2GP models become available — image models in **Image Studio**, video models reachable via the same generation API (Image Studio rejects video output explicitly; full Video Studio wiring is on the roadmap).\n\n| Model | Type | Notes |\n|---|---|---|\n| **Flux.1 Dev** | Image | 1024px, 28 steps |\n| **Qwen Image** | Image | 1024px, 30 steps |\n| **Wan 2.2 (T2V \u002F I2V)** | Video | Slow on consumer GPUs |\n| **Hunyuan Video** | Video | High-quality T2V |\n| **LTX Video** | Video | Fastest video option |\n\n> **Why a separate server?** Wan2GP's runtime (Sage attention, flash-attn, AWQ\u002FGGUF kernels) is CUDA-only — there is no MPS \u002F Apple Silicon path. Treating it as a remote server lets a Mac-only user keep the desktop app while offloading inference to a Linux\u002FWindows GPU box, a gaming PC on the LAN, or a rented RunPod\u002Fvast.ai instance.\n\n> **Local inference is only available in the desktop app.** The hosted web version always uses cloud APIs.\n\n### Hardware Notes\n\n- **sd.cpp** runs on CPU (all platforms) and **Metal GPU** on Apple Silicon (M1\u002FM2\u002FM3\u002FM4); CUDA\u002FVulkan\u002FROCm on Linux\u002FWindows.\n- Metal GPU acceleration is built into the macOS desktop binary — significantly faster than CPU-only.\n- Recommended for sd.cpp Z-Image: 16 GB RAM (7.4 GB weights + 2.4 GB compute buffer). On a base 8 GB M-series Mac, **Z-Image is known to hang the system** — stick to SD 1.5 there.\n- For SD 1.5 on M2: expect ~1–2 s\u002Fstep with the Metal dylib active. If you see ~10 s\u002Fstep instead, the binary may have fallen back to CPU — see verification below.\n\n### Verifying the SD 1.5 path (the fastest sanity test on Mac)\n\nIf you want to confirm sd.cpp is installed correctly without going through the UI, you can drive `sd-cli` directly. This is the same binary the app uses.\n\n```bash\n# 1. App data layout (created on first app launch)\nAPP_DATA=\"$HOME\u002FLibrary\u002FApplication Support\u002Fopen-generative-ai\u002Flocal-ai\"\nls \"$APP_DATA\u002Fbin\"     # sd-cli, libstable-diffusion.dylib\nls \"$APP_DATA\u002Fmodels\"  # whatever you've downloaded\n\n# 2. Grab a small SD 1.5 model directly (Dreamshaper 8, ~2 GB)\ncurl -L --fail --progress-bar \\\n  -o \"$APP_DATA\u002Fmodels\u002FDreamShaper_8_pruned.safetensors\" \\\n  \"https:\u002F\u002Fhuggingface.co\u002FLykon\u002FDreamShaper\u002Fresolve\u002Fmain\u002FDreamShaper_8_pruned.safetensors\"\n\n# 3. Run a single 512x512 \u002F 12-step inference\nDYLD_LIBRARY_PATH=\"$APP_DATA\u002Fbin\" \"$APP_DATA\u002Fbin\u002Fsd-cli\" \\\n  -m \"$APP_DATA\u002Fmodels\u002FDreamShaper_8_pruned.safetensors\" \\\n  -p \"a serene mountain lake at sunrise, oil painting\" \\\n  -o \u002Ftmp\u002Fsd15-test.png \\\n  --steps 12 -H 512 -W 512 --cfg-scale 7.5 --seed 42 \\\n  --sampling-method euler_a\n```\n\nA healthy run on Apple Silicon prints `total params memory size = 1969.78MB (VRAM 1969.78MB, RAM 0.00MB)` (Metal-backed) and produces a coherent 512×512 PNG. If `VRAM` is `0.00MB` instead, the dylib is CPU-only — check `otool -L \"$APP_DATA\u002Fbin\u002Flibstable-diffusion.dylib\" | grep -i metal` and reinstall the engine from **Settings → Local Models** if Metal is missing.\n\n---\n\n## ✨ Features\n\n- **Image Studio** — Generate images from text prompts (50+ text-to-image models) or transform existing images (55+ image-to-image models). Switches model set automatically based on whether a reference image is provided. Quality and resolution controls visible for models that support them.\n- **Local Inference** — Two engines: **sd.cpp** (bundled, runs on Mac\u002FWin\u002FLinux with Metal\u002FCUDA\u002FVulkan\u002FROCm) for SD 1.5, SDXL, and Z-Image; and **Wan2GP** (BYO Gradio server) for Flux, Qwen-Image, and video models (Wan 2.2, Hunyuan, LTX). Configure both in Settings → Local Models.\n- **Multi-Image Input** — Upload up to 14 reference images for compatible edit models (Nano Banana 2 Edit, Flux Kontext Dev, GPT-4o Edit, and more). Multi-select picker with order badges, batch upload, and a \"Use Selected\" confirmation flow.\n- **Video Studio** — Generate videos from text prompts (40+ text-to-video models) or animate a start-frame image (60+ image-to-video models). Same intelligent mode switching as Image Studio.\n- **Lip Sync Studio** — Animate portrait images or sync lips on existing videos using audio. 9 dedicated models across two modes: portrait image + audio → talking video, and video + audio → lipsync video.\n- **Cinema Studio** — Interface for photorealistic cinematic shots with pro camera controls (Lens, Focal Length, Aperture)\n- **Workflow Studio** — Build and run multi-step AI pipelines visually. Chain image, video, and audio models into automated flows. Browse community templates, create your own with a node-based editor, and run them via an interactive playground.\n- **Upload History** — Reference images are uploaded once and stored locally. A picker panel lets you reuse any previously uploaded image across sessions — no re-uploading.\n- **Smart Controls** — Dynamic aspect ratio, resolution\u002Fquality, and duration pickers that adapt to each model's capabilities (including t2i models with resolution or quality options)\n- **Generation History** — Browse, revisit, and download all past generations (persisted in browser storage)\n- **Image & Video Download** — One-click download of generated outputs in full resolution\n- **API Key Management** — Secure API key storage in browser localStorage (never sent to any server except Muapi)\n- **Responsive Design** — Works seamlessly on desktop and mobile with dark glassmorphism UI\n\n### 🖼️ Image Studio — Dual Mode\n\nThe Image Studio automatically switches between two model sets:\n\n| Mode | Trigger | Models | Prompt |\n| :--- | :--- | :--- | :--- |\n| **Text-to-Image** | Default (no image) | 50+ t2i models (Flux, Nano Banana 2, Seedream 5.0, Ideogram, GPT-4o, Midjourney…) | Required |\n| **Image-to-Image** | Reference image uploaded | 55+ i2i models (Kontext, Nano Banana 2 Edit, Seedream 5.0 Edit, Seededit, Upscaler…) | Optional |\n\n#### Newly Added Models\n\n| Model | Type | Key Features |\n| :--- | :--- | :--- |\n| **Nano Banana 2** | Text-to-Image | Google Gemini 3.1 Flash Image · Resolution 1K\u002F2K\u002F4K · Google Search enhancement · aspect ratio `auto` |\n| **Nano Banana 2 Edit** | Image-to-Image | Up to **14 reference images** · Resolution 1K\u002F2K\u002F4K · Google Search enhancement |\n| **Seedream 5.0** | Text-to-Image | ByteDance · Quality basic\u002Fhigh · 8 aspect ratios · up to 4K |\n| **Seedream 5.0 Edit** | Image-to-Image | ByteDance · Natural language style transfer · Quality basic\u002Fhigh |\n| **MiniMax Image 01** | Text-to-Image | MiniMax · 8 aspect ratios · up to 4 images per request · 1500 char prompt |\n\n#### Multi-Image Input\n\nModels that accept multiple reference images expose a multi-select picker when active:\n\n| Model | Max Images |\n| :--- | :--- |\n| Nano Banana 2 Edit | 14 |\n| Nano Banana Edit | 10 |\n| Flux Kontext Dev I2I | 10 |\n| Kling O1 Edit Image | 10 |\n| GPT-4o Edit \u002F GPT Image 1.5 Edit | 10 |\n| Bytedance Seedream Edit v4 \u002F v4.5 | 10 |\n| Vidu Q2 Reference to Image | 7 |\n| Flux 2 Flex\u002FPro Edit | 8 |\n| Nano Banana Pro Edit | 8 |\n| Flux Kontext Pro\u002FMax I2I | 2 |\n| Wan 2.5\u002F2.6 Image Edit | 2–3 |\n| Qwen Image Edit Plus \u002F 2511 | 3 |\n| GPT-4o Image to Image | 5 |\n| Flux 2 Klein 4b\u002F9b Edit | 4 |\n\nWhen a multi-image model is selected the upload trigger switches to multi-select mode:\n- **Checkboxes with order numbers** — images are sent to the model in the order you select them\n- **Batch upload** — pick multiple files at once from your file dialog\n- **Count badge** on the trigger shows how many images are active; a `+` badge appears when more slots are available\n- **\"Use Selected\" button** confirms and closes the picker\n\n### 🎬 Video Studio — Dual Mode\n\nThe Video Studio follows the same pattern:\n\n| Mode | Trigger | Models | Prompt |\n| :--- | :--- | :--- | :--- |\n| **Text-to-Video** | Default (no image) | 40+ t2v models (Kling, Sora, Veo, Wan, Seedance 2.0, Hailuo, Runway…) | Required |\n| **Image-to-Video** | Start frame uploaded | 60+ i2v models (Kling I2V, Veo3 I2V, Runway I2V, Wan I2V, Seedance 2.0 I2V, Midjourney I2V…) | Optional |\n\n#### Newly Added Models\n\n| Model | Type | Key Features |\n| :--- | :--- | :--- |\n| **Seedance 2.0** | Text-to-Video | ByteDance · Aspect ratios 16:9 \u002F 9:16 \u002F 4:3 \u002F 3:4 · Duration 5 \u002F 10 \u002F 15s · Quality basic\u002Fhigh |\n| **Seedance 2.0 I2V** | Image-to-Video | ByteDance · Animate images into video · Up to 9 reference images · Aspect ratios 16:9 \u002F 9:16 \u002F 4:3 \u002F 3:4 · Duration 5 \u002F 10 \u002F 15s · Quality basic\u002Fhigh |\n| **Seedance 2.0 Extend** | Video Extension | ByteDance · Seamlessly continue any Seedance 2.0 generation · Preserves style, motion & audio · Optional continuation prompt · Duration 5 \u002F 10 \u002F 15s · Quality basic\u002Fhigh |\n| **Grok Imagine T2V** | Text-to-Video | xAI · Duration 6 \u002F 10 \u002F **15s** · Modes: fun \u002F normal \u002F spicy · Aspect ratios 9:16 \u002F 16:9 \u002F 2:3 \u002F 3:2 \u002F 1:1 |\n| **Grok Imagine I2V** | Image-to-Video | xAI · Duration 6 \u002F 10 \u002F **15s** · Modes: fun \u002F normal \u002F spicy · Cinematic motion from still images |\n| **MiniMax Hailuo 02 \u002F 2.3 Standard & Pro** | Text-to-Video \u002F Image-to-Video | MiniMax · Full HD video · Multiple aspect ratios · Fast variant included |\n\n### 🎙️ Lip Sync Studio\n\nThe **Lip Sync Studio** generates audio-driven talking videos using 9 models across two input modes:\n\n| Mode | Trigger | Description |\n| :--- | :--- | :--- |\n| **Portrait Image** | Default | Upload a portrait image + audio file → animated talking video |\n| **Video** | Switch to Video mode | Upload an existing video + audio file → lipsync video |\n\n#### Image-based Models (Portrait Image + Audio → Video)\n\n| Model | Endpoint | Resolutions | Prompt |\n| :--- | :--- | :--- | :--- |\n| **Infinite Talk** | `infinitetalk-image-to-video` | 480p, 720p | Optional |\n| **Wan 2.2 Speech to Video** | `wan2.2-speech-to-video` | 480p, 720p | Optional |\n| **LTX 2.3 Lipsync** | `ltx-2.3-lipsync` | 480p, 720p, 1080p | Optional |\n| **LTX 2 19B Lipsync** | `ltx-2-19b-lipsync` | 480p, 720p, 1080p | Optional |\n\n#### Video-based Models (Video + Audio → Lipsync Video)\n\n| Model | Endpoint | Resolutions | Prompt |\n| :--- | :--- | :--- | :--- |\n| **Sync Lipsync** | `sync-lipsync` | — | — |\n| **LatentSync** | `latentsync-video` | — | — |\n| **Creatify Lipsync** | `creatify-lipsync` | — | — |\n| **Veed Lipsync** | `veed-lipsync` | — | — |\n| **Infinite Talk V2V** | `infinitetalk-video-to-video` | 480p, 720p | Optional |\n\n**How it works:**\n1. Select **Portrait Image** or **Video** mode using the toggle\n2. Upload your portrait image (or video) using the image\u002Fvideo upload button\n3. Upload your audio file using the audio upload button\n4. Optionally enter a prompt to guide the motion style\n5. Select a model and resolution (where supported), then click **Generate**\n\nGeneration history is saved separately in `lipsync_history` and pending jobs resume automatically on page reload.\n\n### 🔀 Workflow Studio\n\nThe **Workflow Studio** lets you build and run multi-step AI pipelines without writing code.\n\n**Key capabilities:**\n- **Templates** — Start from pre-built workflows (image chains, video pipelines, and more)\n- **My Workflows** — Save and manage your own custom pipelines\n- **Community** — Browse and run workflows published by other users\n- **Node-based Builder** — Drag-and-drop visual editor to connect models and route outputs between steps\n- **Playground** — Run any workflow interactively with a form UI; results render inline\n- **API execution** — Every workflow is also callable via the Muapi API\n\n> 💡 **Want to add workflows to your own app?** Check out **[Vibe Workflow](https:\u002F\u002Fgithub.com\u002FSamurAIGPT\u002FVibe-Workflow)** — the open-source workflow engine powering this feature. Drop it into any project.\n\n### 🎥 Cinema Studio Controls\n\nThe **Cinema Studio** offers precise control over the virtual camera, translating your choices into optimized prompt modifiers:\n\n| Category | Available Options |\n| :--- | :--- |\n| **Cameras** | Modular 8K Digital, Full-Frame Cine Digital, Grand Format 70mm Film, Studio Digital S35, Classic 16mm Film, Premium Large Format Digital |\n| **Lenses** | Creative Tilt, Compact Anamorphic, Extreme Macro, 70s Cinema Prime, Classic Anamorphic, Premium Modern Prime, Warm Cinema Prime, Swirl Bokeh Portrait, Vintage Prime, Halation Diffusion, Clinical Sharp Prime |\n| **Focal Lengths** | 8mm (Ultra-Wide), 14mm, 24mm, 35mm (Human Eye), 50mm (Portrait), 85mm (Tight Portrait) |\n| **Apertures** | f\u002F1.4 (Shallow DoF), f\u002F4 (Balanced), f\u002F11 (Deep Focus) |\n\n### 📁 Upload History & Picker\n\nEvery image you upload is saved locally (URL + thumbnail) so you never upload the same file twice:\n\n- Click the upload button to open the **reference image picker**\n- Previously uploaded images appear in a 3-column grid with thumbnails\n- **Single-image models** — click a thumbnail to instantly select and close\n- **Multi-image models** — toggle multiple thumbnails (shown with order numbers), then click **Use Selected**\n- Upload new images with the **Upload files** button (supports multi-file selection in multi-image mode)\n- Remove individual images from history with the ✕ button\n- History persists across browser sessions (stored in `localStorage`)\n\n## 🚀 Quick Start\n\n### Prerequisites\n\n- [Node.js](https:\u002F\u002Fnodejs.org\u002F) (v18+)\n- A [Muapi.ai](https:\u002F\u002Fmuapi.ai) API key\n\n### Setup\n\n> **Most users want the desktop app, not this dev path.** If you just want to run Open Generative AI on your machine, [download a prebuilt installer](#-download-desktop-app) instead — no Node.js required. The instructions below are for contributors building from source.\n\nPick the entry point that matches your goal:\n\n- **Desktop app (Electron)** → `npm run electron:dev`\n- **Hosted web version (Next.js)** → `npm run dev`\n\n```bash\n# Clone the repository (with submodules — required for the workflow + agent packages)\ngit clone --recurse-submodules https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-Generative-AI.git\ncd Open-Generative-AI\n\n# If you already cloned without --recurse-submodules, run this once:\n# git submodule update --init --recursive\n\n# Install dependencies + build workspace packages (studio, workflow, agents).\n# This step is REQUIRED — `npm install` alone is not enough; the workspaces\n# need to be built before either dev script will work.\nnpm run setup\n\n# Then start ONE of:\nnpm run electron:dev   # Desktop app (Electron + Vite) — recommended\nnpm run dev            # Hosted web version (Next.js) → http:\u002F\u002Flocalhost:3000\n```\n\nYou'll be prompted to enter your Muapi API key on first use (skip the key if you only plan to use local models).\n\n> **Troubleshooting — `Couldn't find a 'pages' directory`**: this means Next.js can't see the `app\u002F` folder. Confirm you're running `npm run dev` from the repo root (the directory that contains `app\u002F`, `package.json`, and `next.config.mjs`), and that you cloned with submodules. Re-run `npm run setup` if `packages\u002FVibe-Workflow` or `packages\u002FOpen-Poe-AI` are empty.\n\n### Production Build\n\n```bash\nnpm run build\nnpm run start\n```\n\n### Desktop App Build\n\nBuild native desktop apps with Electron:\n\n```bash\n# macOS (DMG — Intel + Apple Silicon)\nnpm run electron:build\n\n# Windows (NSIS installer — x64 + ARM64)\nnpm run electron:build:win\n\n# Linux (AppImage + DEB — x64)\nnpm run electron:build:linux\n\n# Both platforms in one pass\nnpm run electron:build:all\n```\n\nInstallers are output to the `release\u002F` folder. Pre-built binaries are also available on the [Releases page](https:\u002F\u002Fgithub.com\u002FAnil-matcha\u002FOpen-Generative-AI\u002Freleases).\n\n## 🏗️ Architecture\n\nThe app is a **Next.js monorepo** with a shared `packages\u002Fstudio` component library.\n\n```\nOpen-Generative-AI\u002F\n├── app\u002F                        # Next.js App Router\n│   ├── layout.js               # Root layout (Tailwind, fonts)\n│   ├── page.js                 # Redirects → \u002Fstudio\n│   └── studio\u002F\n│       └── page.js             # Studio page — renders StandaloneShell\n├── components\u002F\n│   ├── StandaloneShell.js      # Tab nav + BYOK (API key from localStorage)\n│   └── ApiKeyModal.js          # API key entry modal\n├── packages\u002F\n│   └── studio\u002F                 # Shared React component library\n│       └── src\u002F\n│           ├── index.js        # Exports: ImageStudio, VideoStudio, LipSyncStudio, CinemaStudio, WorkflowStudio\n│           ├── models.js       # 200+ model definitions (single source of truth)\n│           ├── muapi.js        # API client (named exports, apiKey as first param)\n│           └── components\u002F\n│               ├── ImageStudio.jsx    # Dual-mode t2i\u002Fi2i studio\n│               ├── VideoStudio.jsx    # Dual-mode t2v\u002Fi2v studio\n│               ├── LipSyncStudio.jsx  # Portrait\u002Fvideo + audio → talking video\n│               ├── CinemaStudio.jsx   # Pro studio with camera controls\n│               └── WorkflowStudio.jsx # Multi-step pipeline builder & playground\n├── next.config.mjs             # transpilePackages: ['studio']\n├── tailwind.config.js\n└── package.json                # workspaces: [\"packages\u002Fstudio\"]\n```\n\nThe `packages\u002Fstudio` library is also consumed by the hosted version on [muapi.ai](https:\u002F\u002Fmuapi.ai) — model updates made in `packages\u002Fstudio\u002Fsrc\u002Fmodels.js` apply to both the self-hosted app and the hosted version automatically.\n\n## 🔌 API Integration\n\nThe app communicates with [Muapi.ai](https:\u002F\u002Fmuapi.ai) using a two-step pattern:\n\n1. **Submit** — `POST \u002Fapi\u002Fv1\u002F{model-endpoint}` with prompt and parameters\n2. **Poll** — `GET \u002Fapi\u002Fv1\u002Fpredictions\u002F{request_id}\u002Fresult` until status is `completed`\n\nAuthentication uses the `x-api-key` header. During development, a Vite proxy handles CORS by routing `\u002Fapi` requests to `https:\u002F\u002Fapi.muapi.ai`.\n\nFile uploads use `POST \u002Fapi\u002Fv1\u002Fupload_file` (multipart\u002Fform-data) and return a hosted URL that is passed to image-conditioned models. For multi-image models the full `images_list` array is forwarded to the API in one request.\n\nLip sync jobs use the same two-step pattern: a dedicated `processLipSync()` method accepts `image_url` or `video_url` alongside `audio_url`, dispatches to the model's endpoint, and polls until the output video URL is available.\n\n## 🎨 Supported Model Categories\n\n| Category | Count | Examples |\n|---|---|---|\n| **Text-to-Image** | 50+ | Flux Dev, Nano Banana 2, Seedream 5.0, Ideogram v3, Midjourney v7, GPT-4o, SDXL |\n| **Image-to-Image** | 55+ | Nano Banana 2 Edit (×14), Flux Kontext Pro, GPT-4o Edit, Seededit v3, Upscaler, Background Remover |\n| **Text-to-Video** | 40+ | Kling v3, Sora 2, Veo 3, Wan 2.6, Seedance 2.0, Seedance 2.0 Extend, Seedance Pro, Hailuo 2.3, Runway Gen-3 |\n| **Image-to-Video** | 60+ | Kling v2.1 I2V, Veo3 I2V, Runway I2V, Seedance 2.0 I2V, Midjourney v7 I2V, Hunyuan I2V, Wan2.2 I2V |\n| **Lip Sync** | 9 | Infinite Talk I2V, Wan 2.2 Speech to Video, LTX 2.3 Lipsync, LTX 2 19B Lipsync, Sync, LatentSync, Creatify, Veed, Infinite Talk V2V |\n\n## 🛠️ Tech Stack\n\n- **Next.js 14** — App Router, server components, fast dev server\n- **React 18** — Studio UI components\n- **Tailwind CSS v3** — Utility-first styling\n- **npm workspaces** — Monorepo with shared `packages\u002Fstudio` library\n- **Muapi.ai** — AI model API gateway\n\n## 🤔 How is this different from other AI Video Plaforms?\n\n**Open Generative AI** is a community-driven, open-source alternative that provides similar creative capabilities without the closed ecosystem:\n\n| | Other providers | Open Generative AI |\n| :--- | :--- | :--- |\n| **Cost** | Subscription-based | Free (open-source) |\n| **Content filters** | Yes — prompts blocked or altered | None |\n| **Restrictions** | Platform guardrails enforced | Full creative freedom |\n| **Models** | Proprietary | 200+ open & commercial models |\n| **Multi-image input** | Limited | Up to 14 images per request |\n| **Lip sync** | No | 9 models, image & video modes |\n| **Hosted version** | Subscription | Free at [muapi.ai\u002Fopen-generative-ai](https:\u002F\u002Fmuapi.ai\u002Fopen-generative-ai) |\n| **Self-hosting** | No | Yes |\n| **Customizable** | No | Fully hackable |\n| **Data privacy** | Cloud-based | Your data stays local |\n| **Source code** | Closed | MIT licensed |\n\n## 📄 License\n\nMIT\n\n## 🙏 Credits\n\nBuilt with [Muapi.ai](https:\u002F\u002Fmuapi.ai) — the unified API for AI image and video generation models.\n\n---\n**Deep Dive**: For more details on the \"AI Influencer\" engine, upcoming \"Popcorn\" storyboarding features, and the future of this project, read the [full technical overview](https:\u002F\u002Fmedium.com\u002F@anilmatcha\u002F).\n\n---\n*Looking for a free, open-source AI Video Platform? Open Generative AI is an open-source AI image and video generation studio — with no content filters that you can self-host, customize, and extend.*\n","Open Generative AI 是一个开源的AI图像和视频生成平台，提供了200多种模型（如Flux、Midjourney、Kling、Sora、Veo）来生成无限制的内容。其核心功能包括无内容过滤器、自托管、MIT许可，用户可以自由地生成和编辑AI图像与视频。此外，该项目支持通过终端自动化媒体生成流程，并提供在线和桌面应用程序版本，无需安装Node.js即可使用。适用于需要高度自由度和定制化能力的创意工作者，以及对现有AI视频平台有替代需求的用户。",2,"2026-06-11 03:32:04","top_language"]