If something is published on the open web, you can drop it into Paper as a source. Paper renders the cleaned article into a PDF so it behaves like any other PDF source - searchable, citeable, summarisable.
Add a web source
- In a page, click + Add source → Paste a link and paste the URL.
- Paper fetches the page, strips navigation and ads, and saves the readable text and title.
- A PDF is generated from the extracted article and indexed.
Behind the scenes Paper tries a chain of extraction methods so paywalled and JS-rendered pages still work where possible:
- Direct fetch with Mozilla Readability (fast path, works for most articles).
- Google Web Cache - often has the full text of paywalled news articles.
- Archive.org Wayback Machine - for pages that have a snapshot.
- Jina Reader - renders JavaScript-heavy pages in a headless browser.
Whichever method returns the most content wins. URL extraction is rate-limited to 10 pages per minute per user.
What works well
- Long-form articles (blogs, news features, op-eds)
- Wikipedia and reference sites
- Online textbook chapters that aren't gated behind a hard paywall
- Documentation and how-to guides
What doesn't
- Pages behind a login - Paper sees the page as a logged-out visitor. The fallback chain helps with soft paywalls but not authenticated content.
- Very short pages - if the article is under ~200 characters of body text after cleaning, Paper rejects it.
- Video-heavy pages - Paper extracts text only. Use the YouTube/video flow for the video itself.
Multiple articles in one page
Adding several web sources to the same page means chat reasons across all of them. Useful for essay research, comparing sources, or building an argument from a collection of articles.