Theme
Language

One paper in.
A poster, a video, a blog out.

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.

🖼️ Live demo Make your own poster Turn any paper into a unified artifact — try it now
100Benchmark papers
5Composable skills
3Artifacts / paper
3.56 vs 3.03Poster aesthetics vs authors
74–95/100Wins vs author posters

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.

Five skills, one pipeline

extract once · generate three ways · converge

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.

1
📄

paper2assets

Reads the PDF once into one reusable bundle — body text, cleaned figures, captions, metadata, and a nine-section spec that every generator consumes.

2Generate — in parallel
🖼️

paper2poster

Composes a print-ready poster in four coupled formats — HTML, PDF, PNG, and editable PowerPoint — tuned by a measured fill-loop.

🎬

paper2video

Plans narration and builds a captioned talk video with an editable deck and narration-aligned visual highlights.

✍️

paper2blog

Writes a bilingual editorial package — a Chinese WeChat piece and an English research blog — as editable Word files sharing one evidence base.

3
🎞️

paper2reel

Binds poster regions, video segments, slide thumbnails, captions, and blog passages into one interactive, downloadable viewer.

paper2assets paper2poster·paper2video·paper2blog paper2reel
Loading reel…