A frontend that wrangles AI into a storyboard pipeline
Print Film forces generative video into a structured workflow—script, assets, keyframes, export—instead of praying to the prompt gods.

What it does Print Film is a React + Electron workbench that turns a story idea into short drama or motion-comic clips. You write a plot, generate character and scene assets, produce start/end keyframes for each shot, and let a video model interpolate the motion. Everything lives in the browser (IndexedDB) or ships as a Docker container on port 3005.
The interesting bit The project treats AI like a production pipeline, not a slot machine. By locking down character design sheets and scene concept art first, then generating keyframes before video, it tries to solve the consistency problem that breaks most text-to-video experiments. The “director’s workbench” even lets you split long shots into sub-shots for finer timing control.
Key highlights
- Four-phase workflow: Script → Assets → Keyframes → Export
- Character “design sheets” with costume variants (daily, combat, formal, injured)
- Keyframe-driven generation: start and end frames constrain the video model’s interpolation
- Desktop build via Electron; Docker + Nginx for self-hosting
- All project data stays local in IndexedDB; only the GitCC API key leaves the machine
- Supports OpenAI-style API endpoints, so you can point it at compatible providers
Caveats
- The README is bilingual but rough; some sections repeat and links are bare URLs
- No backend of its own—if you wipe browser data, your project vanishes unless you exported
- Hard dependency on GitCC API (or a compatible drop-in) for text, image, and video generation
Verdict Worth a spin if you’re producing short vertical dramas or motion comics and tired of manually stitching inconsistent AI clips. Skip it if you need real-time collaboration or a backend that persists your work.