A bookmark library that stays out of your way.

Privacy Policy

No sign-up, no account

Bookmar has no login, no user accounts, and no user identity of any kind. There is nothing to sign up for and nothing tying your activity to you.

You control where your data lives

Bookmar has no backend server of its own — it's a static app with no API to send your data to. Today, your folders and bookmarks (URL, title, description, cover image, favicon, and the date you added them) are stored entirely in your browser's local storage, on your own device, and nothing is synced, backed up, or transmitted anywhere.

• As the Chrome extension, this storage is scoped to the extension itself (chrome-extension://<id>) — it's isolated from every website you visit and from any other version of Bookmar you might run (e.g. a local dev build).
• Bookmar's storage layer is built to support additional storage providers in the future — for example, keeping your library in your own cloud drive instead of only on-device. Whichever provider your library lives in is always something you choose and can see in Settings; it never routes through a server Bookmar operates. This policy will be updated before any such option ships.

What we don't collect

Bookmar contains no analytics, no crash/error reporting, no advertising SDKs, and no cookies. We don't know how many people use Bookmar, what they bookmark, or how they use it — we have no mechanism to find out.

The only network requests Bookmar makes

Saving, editing, and organizing bookmarks all work fully offline — adding a link is instant and never waits on the network. Separately, right after you add a bookmark, Bookmar also attempts to look up its title, description, cover image, and icon so your library looks good; if that lookup can't complete (e.g. no connection at the time), the bookmark still gets a usable fallback title instead of being blocked or retried silently in the background. This lookup is the only reason Bookmar talks to the internet, and only the URL you're bookmarking is sent — never your other bookmarks, folders, or any identifying information.

How the lookup actually works, in order:

1. A few fast, lightweight services are tried in parallel, not one-by-one: they're each asked for a title, cover image, and (for one of them) a description, and whichever answers first with a usable piece of information wins that piece. None of these involve Bookmar reading the bookmarked page's actual content — they're asked about the URL, and answer with a summary.
2. Only if a cover image or description is still missing after that does Bookmar go a step further and read the bookmarked page's own HTML directly, extracting whatever og:/twitter:/<title>/favicon tags it finds. This is the one step where more than just the URL leaves your device, since fetching a page necessarily requests its full content — same as opening it in a browser tab would.
3. If nothing above returns anything, Bookmar falls back to a hostname-derived title and a generic favicon, so the bookmark still looks presentable — no further requests are made.

The services involved:

• noembed.com — a general-purpose lookup used for most sites.
• tiktok.com/oembed and youtube.com/oembed — the official oEmbed endpoints for those two platforms specifically, used instead of the general lookup for better results.
• r.jina.ai — reads the bookmarked page's content to help extract an accurate title when the above don't return one.
• api.microlink.io — a server-side page-metadata lookup, raced alongside the others above; it's the only one of these that can return a description, not just a title/image.
• google.com/s2/favicons — given only the site's domain (not the full URL), returns a favicon when the page doesn't provide its own.
• api.codetabs.com — used only when running Bookmar as a regular web page (not the installed extension), as a fallback to read a page's content, since browsers block a web page from reading another site's content directly. The installed extension doesn't need this: it's able to read the bookmarked page directly, the same way your browser does when you visit it yourself.

None of these requests carry cookies, sign-in credentials, or any Bookmar identifier — they're anonymous, one-off lookups of a single URL.

Chrome extension permissions, explained

• Host permissions (http://*/*, https://*/*) — let the extension fetch the page you're bookmarking directly, on your own device, to pull its title/description/image. This is the same access your browser already has when you open the page yourself; Bookmar doesn't route it through any server of ours.
• tabs — lets the "Add all tabs" button read the URLs of your currently open tabs, so you can bookmark them in bulk. This list is read once, locally, when you click the button, and is only ever used to fill in the add-bookmark form for you to review before saving. It's never sent anywhere.

Changes to this policy

If Bookmar's data handling ever changes, this policy will be updated to reflect it before that change ships.

Contact

Questions about this policy: [add contact email]