Why cluster keywords
A keyword list is not a list of pages. Many keywords mean the same thing to the person searching, and Google often answers them with the same results, so writing a separate page for each one splits your effort and leaves your own pages competing with each other. Grouping the list first turns it into a plan: one page per group, written for its biggest keyword, with the rest worked into its headings and copy.
How it groups them
The keywords are sorted by their search numbers, when you paste them, and each one either joins the group it is closest to in meaning or starts a group of its own. Meaning is measured by an AI embedding model, so “cheap flights to Lagos” and “Lagos flight deals” land together although they share only one word. Keywords that plainly want different things, such as “buy” and “review” searches, or a how-to question and a product search, are never grouped together, since they need different pages, and every group is labelled with its search intent and the kind of page that serves it.
The model runs on Cloudflare’s own servers (Workers AI), and only the keywords you paste are sent to it. Cloudflare does not use what it receives to train models.
How it differs from tools built on paid data
Tools that pay for search data group keywords by the Google results they share, which is the surest test that one page can rank for them all. That data costs money for every keyword, so this tool groups by meaning instead. Close meaning usually means shared results, but not always, so search the main keyword of your biggest groups and check that the top results overlap before you merge pages.
Your keywords and your data
The tool is free. It asks for your email address before it runs, and you only join the mailing list if you tick the box. Your keywords are kept in your report, which has its own private link that works for 90 days and is then deleted. The privacy policy explains the rest.