Paper2Poster
About this tool
The open-source implementation of a NeurIPS 2025 Datasets and Benchmarks paper. PosterAgent extracts text and figures from a paper PDF, plans a layout, renders it and uses vision-model feedback to fix overflow and alignment. It exports editable PPTX posters with configurable dimensions and styles. The repository also includes PaperQuiz evaluation and a lightweight skill.
- Feature 1: Parser, Planner and Painter-Commenter stages with visual feedback.
- Feature 2: Editable poster.pptx output, configurable dimensions such as 48×36 inches and YAML styles.
- Feature 3: Automatic conference and institution logos, with parallel section generation.
- Feature 4: OpenAI models or self-hosted open-source models such as Qwen through vLLM.
- Feature 5: A lightweight skills/ entry helps Claude and Codex prepare poster text and layouts.
Resource information
- Maintainer
- Paper authors from the University of Waterloo, Vector Institute and University of Oxford; GitHub organization Paper2Poster.
- Type
- Standalone app
- Category
- Scientific figures & presentations
- Runs on
- Agent application
- Who it's for
- Graduate students, postdocs and early-career faculty drafting conference or research posters for refinement in PowerPoint.
- License at time of listing
- MIT
- Source
- GitHub
