AI film creation studio — plan, generate, and assemble a film in one place.
Write scripts and characters as AI-assisted documents, generate shots and scenes through a node-based flow builder, pull from music and voice libraries, and edit the results — with generation cost visible at every step.
| Doc | What's inside |
|---|---|
| Product vision | The problem, the vision, core concepts, experience principles, non-goals |
| MVP requirements | Feature requirements for the six MVP areas: Video Generator, Video Editor, Projects, Music Library, Voice Library, Voice Editor |
| Component requirements | The AI components the MVP needs — the bridge to the parallel design-system project |
| Roadmap | Phases: workflow mastery → component library → MVP → beyond |
| Reference products | Flagship AI products per app type — design research feeding the component work |
| POC plan | Implementation plan of the existing storyboard-planner POC |
| System design | Technical design of the POC |
- Docs: vision, MVP requirements, component requirements, and roadmap drafted.
- POC:
film-planner/— a working storyboard planner (paste a screenplay → extracted scenes/shots → AI-generated reference image per shot, with style propagation). It becomes the seed of the MVP's Projects/Documents area.
FilmMaker is built in parallel with an AI design system: a shadcn/ui-based component library extended with AI-specific components (node canvases, generation panels, model pickers, media galleries…). FilmMaker is the discovery vehicle — components proven here are contributed there. See component requirements.
- Next.js 16 + AI SDK v6
- Vercel AI Gateway (gpt-image-1)
- Tailwind v4 + shadcn/ui
- localStorage persistence (POC)
cd film-planner
cp .env.example .env.local # add AI_GATEWAY_API_KEY (or OPENAI_API_KEY)
pnpm dev