Any web page, turned into notes you can revise from
Paste a link to a lecture page, a documentation page or a long article and get structured notes back: headings, key points, definitions and the numbers that matter, with the source wording kept intact.
Reading something once is not the same as having notes on it. This mode fetches the page, keeps the readable text, drops the menus, cookie banners and scripts, and then builds the same structured notes you would get from a PDF.
It works on pages that send their text as HTML: university lecture pages, documentation, Wikipedia, news analysis, blog posts and reports. Pages that sit behind a login, or that draw everything with JavaScript after the page loads, come back nearly empty — paste the text instead and the result is the same.
What the output looks like
Real output: one Wikipedia article in, 15191 characters of notes out, 42.2 seconds. Unedited:
# Testing effect - Wikipedia ## TL;DR The testing effect—also known as retrieval practice, active recall, practice testing, or test-enhanced learning—demonstrates that dedicating time to retrieving information from memory significantly improves long-term retention compared to passive restudy strategies like re-reading or highlighting. Rather than serving merely as assessment tools, tests act as powerful learning events that modify memory and slow the rate of forgetting. Even difficult or unsuccessful retrieval attempts enhance subsequent learning, particularly when paired with feedback. The core takeaway is that memory is strengthened far more by the effortful act of retrieving knowledge fr
Who it is for
- A 6,000-word article you have to remember next week
- Documentation pages you keep re-reading and never write down
- Course or lecture pages where the reading list is the point
Not a good fit: Pages behind a login, and pages that draw their text with JavaScript after loading — those come back nearly empty; paste the text instead.
Three things that improve the result
- Send one page, not a section of a site. Notes on a whole site blur into a summary of nothing.
- If a page fails, try the print or AMP version, or paste the text — the notes are identical.
- Ask for the output language you study in, not the language of the page. Translation happens in the same pass as the notes.
How it works
- Paste the link into the box under the text area.
- Choose the output language — it does not have to match the page.
- Read the notes, copy them, or download them as Markdown.
Questions
Which pages work best?
Pages that carry their text in the HTML: lecture notes, documentation, Wikipedia, long articles and reports. If a page needs a login or is built entirely with JavaScript, the fetch returns almost nothing and you are told so instead of getting invented notes.
Does it follow the links inside the article?
No. It reads the one page you give it, following redirects only. Pages linked from it are separate jobs.
Is the page stored?
No. The text is fetched, used for that one job, and dropped with the rest of the job after about two hours. It is not used to train anything.
Can the notes be in a different language?
Yes — fourteen output languages. The page can be in any language; the notes are written in the one you pick.
What happens with a paywalled article?
Sites that require a subscription usually return a preview or a login wall. You get a clear error rather than a summary of the paywall. Paste the part you can read and it will be turned into notes.