PaperLens

PaperLens turns reading research papers into a conversation. — HIGHLIGHT → EXPLAIN. Select any passage and get a streamed explanation of what it means in the context of THIS paper — its role in the argument, the jargon, the caveats. One click rewrites it Simpler (plainer terms) or Shorter (just the essentials). — PAPER MAP. One click annotates the paper in place — thesis, contributions, methods & validation, benchmarks & results, limitations — color-coded, with jump-to-passage buttons. — ASK ANYTHING. A chat grounded in the paper plus related literature ("Why did they run this ablation? Did other papers do it too?") with linked sources. — RELATED WORK. Finds papers related to the one you're reading, and papers that approach the exact concept you highlighted (via Semantic Scholar). — BRING YOUR OWN AI. Use your Anthropic API key, your OpenAI API key, or — free with a Claude subscription — a small local bridge that uses the Claude Code CLI on your machine. Keys stay in your browser; there is no PaperLens server, no account, no telemetry. Your work persists per paper: highlights, explanations, structure maps and chat history are restored (and re-anchored in the page) when you return. The sidebar reflows the page beside it, so it never covers the text. Works best on arXiv HTML papers (arxiv.org/html/…, ar5iv) and any article-style page. On arxiv.org/abs pages it offers a jump to the full HTML version. Privacy: no analytics, no tracking, no PaperLens server. Selected text and paper text go only to the AI provider YOU configure; paper metadata goes to Semantic Scholar for related-paper lookups. Full policy: see the privacy policy link.