https://www.youtube.com/@PaperExplainAgent
PaperExplainAgent turns a PDF paper into a narrated, Manim-rendered explainer video. The project glues together a simple React UI, a FastAPI job runner, and the TheoremExplainAgent generation pipeline to plan scenes, write Manim code, render clips, and return artifacts back to the browser.
Core idea:
- You upload a research PDF and provide an API key.
- The backend ingests the PDF into markdown + images, summarizes it, and feeds that context to the TheoremExplainAgent pipeline.
- The agent plans scenes, drafts Manim code, renders the scenes, and combines them into a final video (plus intermediate artifacts like logs and thumbnails).
- The frontend polls job status and previews the finished video and artifacts when ready.