A fully self-hosted, open-source AI image, video, cinema, and lip sync studio. Generate AI images and videos using 200+ models — Flux, Midjourney, Kling, Veo, Seedance and more — with zero cloud dependencies.

• Node.js (v18+) — for building the frontend
• Deno (v1.40+) — for the backend + frontend serving

# Clone the repository
git clone git@github.com:virtualcodewarrior/cinema-creator.git
cd cinema-creator

# Install dependencies, build workspace packages, and fetch the sd.cpp binary
npm run setup

# Build the frontend (static export)
npm run build

# Start Deno backend (serves both API and frontend)
deno task start-all

The app will be available at http://localhost:8000. The Deno backend runs in open mode by default — no API key required. To enable API key authentication, set AI_CINEMA_API_KEY or configure it in ~/.ai-cinema/config.json.

npm run setup also downloads the stable-diffusion.cpp engine binary (sd-cli) into ~/.ai-cinema/bin/. It is required for local image generation — without it the backend logs a warning at startup and local models are unavailable.

To (re-)install the binary at any time:

cd deno
deno task fetch-binary

Model weights (2.1–6.9 GB each) are not installed by setup — download the ones you need in Settings → Models.

# Start everything (builds frontend + runs Deno server)
deno task start-all

# Or run Deno backend only (after frontend is built)
deno task dev

Run everything locally with a Deno backend and local AI models. No cloud dependencies, no subscription fees.

# Fetch the sd.cpp engine binary (done automatically by "npm run setup")
cd deno && deno task fetch-binary && cd ..

# Build frontend and start Deno server (serves both API and frontend)
deno task start-all

See deno/README.md for full self-hosted documentation, including how to obtain the sd.cpp binary and model weights manually.

• Image Studio — Generate images from text prompts (50+ text-to-image models) or transform existing images (55+ image-to-image models)
• Video Studio — Generate videos from text prompts (40+ text-to-video models) or animate a start-frame image (60+ image-to-video models)
• Cinema Studio — Interface for photorealistic cinematic shots with pro camera controls (Lens, Focal Length, Aperture)
• Lip Sync Studio — Animate portraits or sync lips to any audio with 9 dedicated models
• Workflow Studio — Build and run multi-step AI pipelines visually. Chain image, video, and audio models into automated flows
• Agent Studio — Multi-turn creative agent that plans and executes generation tasks conversationally
• Design Agent Studio — Canvas-based autonomous design agent for iterative visual work
• Audio Studio — Generate and edit AI audio/music from text prompts
• Marketing Studio — Generate ad and marketing-ready creative variations from a single input
• AI Influencer Studio — Tools for creating and managing consistent AI persona/influencer content
• Smart Controls — Dynamic aspect ratio, resolution/quality, and duration pickers that adapt to each model's capabilities
• Generation History — Browse, revisit, and download all past generations
• Responsive Design — Works seamlessly on desktop and mobile with dark glassmorphism UI

The frontend is a static bundle built with Vite, served by the Deno backend which also handles all API requests.

cinema-creator/
├── src/                        # Vite frontend source
│   ├── main.jsx                # Entry point (React Router)
│   ├── App.jsx                 # App component with routes
│   ├── globals.css             # Tailwind + custom styles
│   └── pages/                  # Page components
├── app/                        # Shared React components
│   ├── app-shell.js            # Unified SPA shell with sidebar navigation
│   └── agents/                 # Agent page components
├── components/
│   ├── SettingsPanel.js        # API key & model settings
│   └── VideoHistoryPanel.js    # Generation history panel
├── deno/                       # Deno backend + frontend server
│   ├── main.ts                 # Entry point (API + static file serving)
│   ├── api/                    # API route handlers
│   ├── inference/              # sd.cpp inference engine
│   ├── lib/                    # Config, auth, queue, dispatcher
│   └── storage/                # File, model, and history storage
├── packages/
│   ├── studio/                 # Shared React component library
│   │   └── src/
│   │       ├── index.js        # Exports all studio components
│   │       ├── models.js       # 200+ model definitions
│   │       ├── backendClient.js # Deno backend API client
│   │       └── components/     # Studio UI components
│   ├── Vibe-Workflow/          # Workflow builder UI + server
│   ├── Open-Poe-AI/            # Agent chat UI + server
│   └── Open-AI-Design-Agent/   # Design agent UI + server
├── out/                        # Built static frontend (generated)
└── package.json                # workspaces for all packages

The frontend communicates with the Deno backend directly via the same origin.

1. Submit — POST /api/v1/{model-endpoint} with prompt and parameters
2. Poll — GET /api/v1/predictions/{request_id}/result until status is completed

Authentication uses the x-api-key header. Since the frontend and API share the same origin, no CORS proxy is needed.

File uploads use POST /api/upload (multipart/form-data) and return a URL that is passed to image-conditioned models.

• Vite — Fast bundler for the frontend
• React 19 — Studio UI components
• React Router — Client-side routing
• Tailwind CSS — Utility-first styling
• Deno — Backend HTTP server + static file serving + job queue + inference dispatcher
• npm workspaces — Monorepo with shared packages

MIT