0h7c8bggs3o0hh72h4fi4_source.mp4 Apr 2026

: Synchronizes a virtual cursor with the narration to highlight specific areas of the slides.

: Automatically generates and refines LaTeX-based slides from the paper's text. 0h7c8bggs3o0hh72h4fi4_source.mp4

: Uses Vision-Language Models (VLMs) to create narration subtitles and visual-focus prompts. : Synchronizes a virtual cursor with the narration

The paper introduces , a multi-agent framework that automatically converts academic papers into professional presentation videos. It breaks the process down into four distinct "builders": 0h7c8bggs3o0hh72h4fi4_source.mp4