Auto-Tagging: Stop Labelling Bookmarks Manually

Tagging is one of the most powerful features in any bookmark manager — and one of the most tedious to do manually. Recallio's auto-tagger solves that by reading the page and suggesting tags for you, no setup required.

How it works

When you save a bookmark, Recallio fetches the page in the background and matches its URL, title, description, and page text against a library of domain rules (github.com → github, code) and keyword patterns (react, docker, machine learning, and dozens more). Within a few seconds, it returns a set of suggested tags.

To be clear about what this is: it's a fast, free, rule-based matcher, not a language model reading for meaning. It won't catch a tag that isn't in its keyword list, and it doesn't understand context the way a human would — but it covers a wide range of common topics well, and it costs nothing to run.

Reviewing auto-generated tags

Suggested tags don't show up on the bookmark card itself — you'll find them when you open a bookmark's edit view. Any suggestion you haven't already added as a real tag appears as a clickable chip; click one to add it to the bookmark's tags. Nothing is applied automatically — you're always the one who decides which suggestions to keep.

What types of content tag well

Auto-tagging works best on pages with real title/description metadata and readable page text: articles, documentation, tutorials, blog posts, GitHub repos. It works less well on:

• Paywalled content — only the public portion gets read.
• Video pages — YouTube and similar pages have minimal text in the page body (though a matching domain still gets tagged video).
• Login-gated pages — the fetcher sees the login wall, not the content.

For videos and gated content, tag manually or use the tag field in the extension popup at the moment of saving — that's when you know what the content is.

Bulk auto-tagging existing bookmarks

Already have a library of untagged bookmarks? Go to Settings → Auto-Tagging. You'll see how many of your bookmarks have suggested tags and how many don't, plus two buttons: Tag untagged bookmarks (fills in suggestions for anything that's never been tagged) and Re-tag entire library (recomputes suggestions for everything, useful if you've imported a lot since the last pass). Both run in the background in small batches, so a large library takes a few minutes rather than locking up the page.

When to trust auto-tags vs. manual tags

Auto-tags are most reliable for:

• Domain and topic classification (react, css, python, finance)
• Content type (tutorial, docs, news, research)

Manual tags are better for:

• Intent (to-read, reference, to-share, review)
• Project association (project-x, client-y)
• Status (active, archived, in-progress)

The best workflow combines both: let the AI handle topic classification, and add one or two intent tags yourself at save time. You get comprehensive tagging with minimal effort.