xeronsh /openOii

About openOii

openOii turns a story idea into characters, shots, video clips and a finished cut. You review it all on an infinite canvas while progress updates live. The author calls it a learning demo, not for production, and it needs API keys. There is no LICENSE file, so reuse rights are unclear.

Best for: Developers who want to study or fork a working multi-agent comic drama pipeline.

• Runs plan, render and compose stages as a LangGraph pipeline (LangGraph chains AI steps together), with an AI review step
• Shows characters, shots and results on a tldraw infinite canvas, a zoomable whiteboard
• Streams live progress; runs can be resumed, cancelled or sent back with feedback
• Scores how consistent each character's face is across shots with InsightFace
• Adds Edge TTS voices and background music, mixed into the video with FFmpeg
• Exports a PDF comic book or a vertical webtoon strip

1. 1
   Clone the repo and open the project folder.
2. 2
   the command below, then fill in your LLM, image and video settings. Set IMAGE_PROVIDER=fake or VIDEO_PROVIDER=fake to test without those APIs.

   cp backend/.env.example backend/.env

   Copy
3. 3
   Start everything with docker-compose up -d
4. 4
   Open http://localhost:15173 (API docs at http://localhost:18765/docs).
5. 5
   For local dev: in backend, run uv sync and the command below; in frontend, run pnpm install and pnpm dev.

   uv run uvicorn app.main:app --reload --host 0.0.0.0 --port 18765

   Copy