opencut v3

AI-driven multi-domain video production

opencut v3 AI-driven multi-domain video production Client · web / mobile · Architecture component Client web / mobile FastAPI + Uvicorn · Python 3.10+ · Architecture component · Pydantic v2 FastAPI + Uvicorn Python 3.10+ Pydantic v2 edge-tts · voice synthesis · Architecture component edge-tts voice synthesis OpenAI SDK · LLM script · Architecture component OpenAI SDK LLM script PyTorch · align extra · Architecture component PyTorch align extra OpenCV · face detect extra · Architecture component OpenCV face detect extra Transformers · align extra · Architecture component Transformers align extra httpx async · external APIs · Architecture component httpx async external APIs Render Pipeline · ffmpeg + compose · Architecture component Render Pipeline ffmpeg + compose REST speak script align detect fetch models compose Legend Backend Cloud External

Plan

  • • FastAPI receives request
  • • OpenAI SDK scripts the cut

Synthesize

  • • edge-tts voices
  • • PyTorch + Transformers align

Render

  • • OpenCV face detect
  • • httpx fetches external media
  • • Render pipeline composes final cut