Link rel attribute auditor: sponsored, ugc, nofollow
Audit the links in pasted HTML against Google's rel qualifiers: sponsored for paid links, ugc for user content and nofollow when neither applies.
Compare titles or H1s across your own pages and flag near-duplicate pairs using Jaccard word-overlap similarity, entirely in your browser.
Runs entirely in your browser. Nothing you enter is uploaded, logged or stored.
Two pages on the same site can end up chasing the same query without anyone deciding that on purpose: a new post gets written, an old guide never gets updated, and both sit in the index with titles that say almost the same thing. This tool compares the titles or H1s you enter and scores how much significant wording they share, so you can spot the pairs worth a closer look before you merge, redirect or rewrite anything.
Each title is lowercased, stripped of punctuation, and split into words. A small, deliberately non-exhaustive English stopword list, words like “a”, “the”, “and”, “of” and “for”, is removed from each side, because these carry no topical signal and would inflate every score. What is left is a set of significant words for each title.
Jaccard similarity is then the size of the intersection of the two sets divided by the size of their union, expressed as a percentage. Two titles that share every significant word score 100%. Two titles that share none score 0%. A pair like “Best running shoes for 2026” and “Top running shoes: 2026 review” shares “running”, “shoes” and “2026” out of a combined set of six significant words, a worked example the tool’s own prefilled rows demonstrate directly.
“Cannibalization” is a term the SEO industry uses, not one Google’s own documentation uses. What Google has actually published, in its guidance on consolidating duplicate URLs, is that when several pages on a site are similar enough, its systems select one canonical URL to show for a given query and may fold the others’ signals into it. Which page gets chosen is decided by Google’s own systems, not by which one you intended to rank.
That is the real, citable mechanism behind the industry term. No Google document states a title similarity percentage that defines when this happens, so treat any number, including the 50% default in this tool, as a heuristic for where to look, not a threshold Google enforces.
Run the check periodically as new pages are published, since overlap tends to build up gradually rather than appear all at once.
No. It is SEO-industry jargon for two pages competing for the same query. Google's own documentation talks about duplicate content and canonicalization instead, where its systems pick one URL to represent a set of near-identical pages in results.
Jaccard similarity on significant words is simple, explainable and fully computable in the browser. It measures literal word overlap in a title, not meaning, so it is a starting signal for a manual check, not a ranking prediction.
No. Google has never published a similarity percentage that defines when two pages compete for a query. The 50% default here is this tool's own configurable heuristic, adjust it to match how strict you want the check to be.
Audit the links in pasted HTML against Google's rel qualifiers: sponsored for paid links, ugc for user content and nofollow when neither applies.
Preview a title and meta description as Google renders them, measured in pixels rather than characters, on desktop and mobile.
Score pasted text with the Flesch reading ease formula and estimate reading time, computed live in your browser from real, cited constants.
Preview how a title, description and image will render as a Facebook or LinkedIn link card and as an X summary_large_image card.