ResearchStudio-Reel automates the last mile of research — turning a single accepted-paper PDF into a print-ready conference poster, a narrated talk video, and a bilingual blog, all bound into one interactive reel.
Each reel is an unedited end-to-end run on the 100-paper Paper2Poster benchmark. A shared extraction pass keeps every poster, video, and blog grounded in the same evidence.
Every tile is a real, unedited run. Hover to enlarge a poster; click to open its video, blog, interactive reel, and summary.
ResearchStudio-Reel runs as five composable skills inside Claude Code and Codex. A shared extractor reads the paper once; three generators produce editor-ready artifacts from that one bundle; a convergence layer binds them into a single navigable surface.
Reads the PDF once into one reusable bundle — body text, cleaned figures, captions, metadata, and a nine-section spec that every generator consumes.
Composes a print-ready poster in four coupled formats — HTML, PDF, PNG, and editable PowerPoint — tuned by a measured fill-loop.
Plans narration and builds a captioned talk video with an editable deck and narration-aligned visual highlights.
Writes a bilingual editorial package — a Chinese WeChat piece and an English research blog — as editable Word files sharing one evidence base.
Binds poster regions, video segments, slide thumbnails, captions, and blog passages into one interactive, downloadable viewer.
paper2assets→
paper2poster·paper2video·paper2blog
→paper2reel