ChatPaper is an open-source research assistant on GitHub, with nearly 20,000 stars, that uses ChatGPT to accelerate the whole academic reading workflow. It downloads and summarizes arXiv papers, translates full PDFs while preserving layout, polishes academic writing, and even drafts peer-review reports and rebuttal responses. A hosted web version runs at chatpaper.org, and the project remains actively maintained, with the repository updated in 2026.
ChatPaper tackles the flood of daily arXiv submissions by reading papers the way a researcher would: it extracts the abstract, introduction, method, and conclusion, then summarizes each section into a fixed, scannable format covering background, prior work, the proposed method, and results. Papers can be pulled automatically by keyword, author, and date filters, or supplied as local PDFs and folders for batch processing. Around the core summarizer, the project bundles companion tools for translation, polishing, review, response writing, and title generation, plus a Flask web interface and Docker deployment for teams.
ChatPaper is free and open source, and its hosted web tools are free to use โ you supply your own OpenAI API key for the self-hosted version and pay only for that usage.
ChatPaper is built for graduate students, researchers, and R&D teams who need to screen large numbers of papers quickly, especially readers working across language barriers. Lab groups can also self-host it as a shared literature-triage service, turning the daily arXiv firehose into a short, readable shortlist for the whole team.