About keepthis.

Put keepthis.link/ in front of any URL in your address bar and press Enter. The page is kept before you land — there is nothing to confirm and no Save button. Its title and preview are fetched, and three to six tags are suggested from the title and description.

You can edit anything, undo it, or just leave. Undo removes the keep and gives you five seconds to change your mind.

A URL fragment — everything after a # — never reaches any server. Browsers keep it on your machine. So keepthis.link/example.com/page#section keeps example.com/page, without the fragment. There is no way around this from a URL alone.

When you keep a page, our server requests it once and reads only the metadata in its head — the title, description, preview image and site name. We do not store the page, screenshot it, or copy its images; we store the image’s URL and load it from the original site.

That result is cached for seven days and shared across everyone, so a page kept by several people is fetched once. The cache holds public page metadata and nothing about who kept it.

Tags come from the page’s title, description and site name, plus the tags you already use — nothing from the page body is ever sent anywhere. The suggestion prefers words already in your vocabulary and only invents one when nothing fits.

If the suggestion is unavailable, tags are derived from the site and the URL instead and labelled as such. Every keep ends up with a title and at least one tag.

Connecting a client gives it three tools: search your keeps, list your recent keeps, and keep a URL. It can read the tags you use.

It sees only your own keeps, and only once you have connected a client and signed in. It cannot see anyone else’s keeps, your browsing history, or any page you have not kept. Disconnect the client and it sees nothing.

There are no collections, folders, sharing, highlights, notes, reader mode, teams or browser extension, and there is no full-text search of page contents. The AI is the reader mode.