Infinite canvas meets AI film studio for solo creators
It wires an infinite whiteboard directly into an AI film pipeline so solo creators can script, storyboard, generate, and rough-cut without tab-hopping.
What it does
影策 is an open-source workbench that treats AI film and short-drama creation like a node-based design problem. It gives you an infinite canvas to arrange scripts, characters, scenes, and storyboards, then dispatches multimodal generation tasks—text, images, video, audio—through a unified async queue. A built-in timeline editor handles clip arrangement, trimming, subtitles, and server-side export, so the loop from idea to rough cut stays inside one browser tab.
The interesting bit
Instead of gluing together a dozen separate generative tools, the project wraps them into structured film workflows—complete with a 3D director’s console and an Agent layer that can summarize the canvas and replay streaming events. There’s even a community skill library with 34 scenario domain packages, suggesting the maintainers want this to behave less like software and more like a film crew that happens to run in a browser.
Key highlights
- Structured film workflows: scripting, casting, scenes, shot lists, and a 3D director’s console, all arranged on a free canvas with nodes, connections, and a minimap.
- Multimodal generation queue: async tasks for text, image, video, and audio, with support for reference images, key frames, camera motion, inpainting, and batch generation.
- Timeline editing: clip arrangement, splitting, trimming, subtitle transcription, and server-side final export with a pluggable editing panel.
- Agent skill ecosystem: persistent dialogue, canvas summaries, streaming replay, plus a community skill repo indexing 34 scenario domain packages.
- Deployment flexibility: SQLite for local solo work, or PostgreSQL/Redis/Docker Compose for server deployment; backend runs Go, frontend runs React/Bun.
Caveats
- The project is in rapid development: data structures and external APIs may shift without warning.
- The cloud Agent is currently read-only; real-world capability should be checked against docs rather than the UI.
- The maintainers explicitly warn against exposing the default build to the public internet for multi-user use without hardening CORS, registration, and upstream model access.
Verdict
Worth a look if you’re experimenting with AI-driven video workflows and want a single surface to storyboard, generate, and rough-cut. If you need a stable, production-grade creative suite or a turnkey SaaS, wait a few releases.
Frequently asked
- What is ddcat-ai/open-ai-canvas?
- It wires an infinite whiteboard directly into an AI film pipeline so solo creators can script, storyboard, generate, and rough-cut without tab-hopping.
- Is open-ai-canvas open source?
- Yes — ddcat-ai/open-ai-canvas is open source, released under the MIT license.
- What language is open-ai-canvas written in?
- ddcat-ai/open-ai-canvas is primarily written in TypeScript.
- How popular is open-ai-canvas?
- ddcat-ai/open-ai-canvas has 1k stars on GitHub.
- Where can I find open-ai-canvas?
- ddcat-ai/open-ai-canvas is on GitHub at https://github.com/ddcat-ai/open-ai-canvas.