YuriyLin/PaperRead

What it solves

Researchers often struggle to move from finding papers to really understanding them. PaperRead combines search, organization, and deep reading in one workspace, so you can go from a research question to annotated PDFs without switching between tools.

How it works

Start by searching a topic or a venue (like CVPR) — PaperRead pulls bibliographic data from OpenAlex (with DBLP and Crossref for venue queries). Save interesting papers to a candidate list; summaries are generated in the background. Then move papers to a reading board where you drag them between statuses such as "to read", "reading", "to summarize", and "done". Every paper card opens a Quick Peek with details. For deep reading, the PDF reader shows the original layout, a translation, your notes, and an AI chat side by side. PaperRead also builds reading paths (ordered recommendations) and mind maps from the paper's content. Summaries and annotations can be exported to Obsidian or as Markdown files.

Who it's for

For researchers, graduate students, and academics who want to manage a large paper collection, read carefully, and take structured notes with AI help. The interface is in Traditional Chinese (an English README is provided). It runs on Windows x64, macOS arm64, and Linux x64.

Highlights

• Research exploration: topic and venue search with source attribution (OpenAlex, DBLP, Crossref)
• Candidate list: saves papers for later, with background summary generation
• Reading board: drag-and-drop status columns, Quick Peek, and filters
• PDF reader: handwriting notes, highlighting, AI explanations, and original-text anchoring
• AI summaries, reading paths, and mind maps grounded in retrieved paper data
• Obsidian vault export for summaries, comparisons, and mind maps
• Multiple AI backends: cloud APIs (OpenAI, Anthropic, Gemini, etc.) and local models (Ollama, LM Studio)
• Desktop apps for Windows, macOS (arm64), Linux AppImage, plus a Docker browser version