Show HN: Paper2Any – Open tool to generate editable PPTs from research papers

Turns research papers into slides — but readers rage at invite-only demo and mixed-language chaos

TLDR: An open tool claims one-click, editable slides and diagrams from research papers. The crowd likes the idea but groans at invite-only access and a mixed Chinese/English README, with top asks for separate readmes and a public demo so researchers can actually try it — adoption hinges on accessibility.

Paper2Any swoops in promising to turn dense research papers into editable slides, diagrams, and posters with one click — dreamy for students and lab folks drowning in PDFs. The devs even tease a slick web demo at dcai-paper2any.cpolar.top and an open-source repo at GitHub. But the community’s reaction? Pure drama. The loudest complaint: the README and demo swing between Chinese and English, leaving some readers locked out. One commenter bluntly asked for two separate readmes — English-only and Chinese-only — so people can actually understand it. Cue the chorus: “Show HN” should be public, not invite-code gated. That “write your own invite_codes.txt” step became instant meme fuel. Others joked it’s the ultimate Death-by-PowerPoint machine, while fans argued it could save hours of figure-making and thesis defense prep. A mini flame war brewed over whether auto-generated diagrams are “helpful summaries” or “pretty nonsense.” The extra setup — Conda, LaTeX, Inkscape — got roasted as “Boss-level unlocks” for what was billed as easy. Still, beneath the snark, there’s excitement: if the team cleans up language and opens the demo, this could be the research-to-slides cheat code everyone secretly wants.

Key Points

  • DataFlow-Agent introduces Paper2Any, a workflow that turns research PDFs/images/text into editable figures, PPTs, videos (planned), and posters (planned).
  • Paper2Figure web beta launched on 2025-12-12 with invite codes; initial project version 0.1.0 released on 2024-09-01.
  • Paper2Figure supports model architecture diagrams, technical route maps (PPTX + SVG, CN/EN), and experimental data charts (under optimization).
  • Setup uses Conda and Python 3.12, with dependencies including Tectonic (LaTeX) and Inkscape; backend runs on FastAPI/uvicorn, frontend on React/Vite.
  • Easy-DataFlow provides AI-driven pipeline recommendation, operator authoring, visual orchestration, prompt optimization, and web data collection; DataFlow-Table is in development.

Hottest takes

"The readme is in large part not in English" — OutOfHere
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.