H1 vs Title Keyword Overlap

The title was written for clicks in the SERP, the H1 was written for someone already on the page, and the two came from different heads. After a few hundred URLs it is common for the main keyword to sit in one and vanish from the other — and whoever reviews the batch has no time to read both side by side, row by row.

The comparison tokenizes both strings, strips accents and punctuation, drops a short stopword list covering English and Portuguese (the, of, for, de, da, para), then computes the Jaccard index: intersection over union of the unique tokens. Bands are 70% or more for strong, 40% for medium, 15% for weak, and anything below that counts as unaligned. Because Jaccard is symmetric, an extra word on either side pulls the score down — the prefilled example scores 29% even though both strings describe the same thing.

Read the number as a signal, not a target. A short title that is a strict subset of the H1 will never approach 100%, and that is fine. There is no stemming: calculator and calculators count as different tokens. The most useful part of the output is the intersection line — if your main keyword is missing from it, the misalignment is real. You paste both strings yourself; the tool never fetches your URL. It runs in the browser.