How it worksTrendingPricingSign in
Home › How it works
How it works

What happens in the 20 seconds after you press "Find the gap".

Theylink is a query on the Linkos backlink index. This page explains what the index is, how gaps are scored, what the labels mean, and where the data stops being certain.

Step 1

Look up both domains

We fetch every page in our index that links to A, and every page that links to B. If a domain is new to us, we crawl its backlinks first — usually two minutes.

Step 2

Subtract

Pages linking to B and not to A become gap rows. Pages linking to both are "shared". Pages linking only to A are shown too — it's not always bad news.

Step 3

Score each row

Linkos Rank of the linking page × topical relevance to your homepage × how easy that page type is to get into × freshness.

Step 4

Render a page

The result gets a permanent URL, an OG image, and a place in our sitemap if both domains have enough data to be worth indexing.

The index

Our own crawl, not someone else's export.

LinkosBot crawls the web and stores every outbound link it finds: source page, target, anchor text, rel attributes, when we first saw it and when we last confirmed it was still there. Pages are rechecked weekly; pages linking to tracked competitors, daily. We store link metadata and page titles — not page bodies.

The index is smaller than Ahrefs' or Semrush's. It's also younger, which means fewer links that were removed in 2021 still counted as live. Every number on a comparison page shows when it was last checked.

Weeklygeneral recheck of every indexed page
Dailypages linking to tracked competitors
24hmaximum age of a cached comparison
# one gap row, simplified score = lr_weight(page.lr) × relevance(page.topic, you.topic) # 0.5 – 1.5 × ease(page.type) × freshness(page.last_checked) ease = { directory: 1.5, # you fill a form listicle: 1.3, # one email to an author review: 1.0, # offer a license blog_mention: 0.8, # a favor forum_ugc: 0.5 # low value, often nofollow } # worked example — the first row on the demo page page = linkos.bio/blog/best-backlink-tools lr_weight(92) → 0.92 relevance("notion alternatives") → 1.30 ease(listicle) → 1.30 freshness(checked 3 hours ago) → 0.98 score = 0.92 × 1.30 × 1.30 × 0.98 → 1.52 # rank 1 of 0
Page types

Why a directory outranks a DR 90 blog mention.

The list is sorted by how likely you are to actually get the link, not by how impressive the domain is. A form you can fill in today beats an editor who might reply next month.

Directory×1.5Alternatives pages, category grids, "tools" lists with a submit button. Detected by URL patterns and page structure. The fix is self-serve.
Listicle×1.3"12 best…", "Top 10…", numbered roundups. Usually one author, usually updated. The fix is a short email with a reason you belong.
Review×1.0A deep dive on one competitor. They test things. The fix is offering a license and a comparison angle.
Blog mention×0.8A passing link in an article. Editors rarely rewrite old posts; low effort to ask, low odds.
Forum / UGC×0.5Reddit, Hacker News, Q&A sites. Often nofollow or ugc. Included for completeness, weighted down.

Classification is rule-based for about 80% of pages. The rest is classified by a language model at crawl time and cached. Mislabels happen; the "report" link on each row fixes it for everyone.

Linkos Rank

LR is our number. It isn't DR.

Linkos Rank is a 0–100 score computed from referring domains in our index, with a weighting pass so a link from a well-linked page counts more than a link from an orphan. It is consistent across every Theylink, Linkrot and Stillinked page. It is not Ahrefs Domain Rating or Moz Domain Authority, and it won't match them. Use it to compare rows on the same page, not to compare with other tools.

What Theylink does not do

  • Sell, place, or broker links
  • Send outreach emails on your behalf
  • Store page bodies or personal data beyond public contact details
  • Index comparisons where either domain has fewer than 20 referring domains
  • Show adult, gambling, or spam domains on public pages
Contacts

Where the editor emails come from, and how to remove one.

On unlocked rows we show a contact for the linking page when we can find one on the site itself: an author byline, an /about or /contact page, or a documented pattern for that domain. Each contact carries a confidence score. We do not scrape LinkedIn or buy lists. Anyone can remove their address at theylink.app/contact-opt-out; removals apply across all Linkos products within 48 hours.

Publicbylines, /about and /contact pages only
Scoredevery contact carries a confidence score
48hto delete after an opt-out is confirmed

That's the whole method. Try it on your competitor.

The comparison is free and takes about twenty seconds.

Compare two domains