Paper2Poster

Why we included this project

Turning a finished paper into a conference poster usually means re-reading the whole thing and hand-placing every figure. Paper2Poster automates that pipeline: it takes a paper PDF, has several agents draft each section in parallel, checks the layout visually as it goes, and hands back an editable .pptx instead of a locked template, so you can still rearrange things by hand before printing. Researchers, PhD students, and lab assistants can run it against hosted models like GPT-4o or local open-weight models served through vLLM, depending on whether they care more about cost or output quality. It also fetches conference and institution logos automatically and accepts YAML style files, which helps a lab producing several posters for one conference keep them consistent without redoing the layout every time.