Keyword clustering

Keyword clustering turns a long keyword list into the pages to write. This tool groups keywords the way Google does: two keywords belong on one page when Google ranks the same pages for both. If “keyword clustering” and “keyword clustering tool” share 4 of their top 10 results, one page can rank for both. If they share none, they need a page each.

Paste your keywords or upload an export, pick the country and language, and run it. Each cluster is named after its biggest keyword and shows its total searches, what its searchers want, every keyword in it, and the pages that already rank for them, so you can see why they belong together.

Sign in to run it. It’s free up to 100 keywords. Bigger lists cost 1 credit ($1) per 1,000 keywords. A run that fails gives its credits back, and I’ll email you when the report is ready (you can turn that off in your account settings).

What you get

• Clusters, biggest first, each named after its biggest keyword, with its total monthly searches, its search intent (learn, compare, use or buy, or find a site) and every keyword in it with its volume.
• The evidence for each cluster: how many of Google’s top 10 results each keyword shares with the biggest one, and the pages that rank for most of them.
• An Ungrouped list: keywords that share too few results with any cluster, each with the reason and the nearest cluster.
• A CSV that adds a cluster column to every keyword, ready to paste back into your sheet, and the whole report as JSON.
• Copy for agent, to plan the pages with your AI agent.

How it works

• It reads Google’s top 10 organic results for every keyword in your country and language, from DataForSEO. Results read for anyone in the last 30 days are reused.
• Keywords are taken biggest first. Each one joins the cluster whose biggest keyword shares the most of its top 10, if they share at least as many results as your strictness. Otherwise it starts a cluster of its own.
• A keyword is only ever compared with a cluster’s biggest keyword, never with the others in it. So a cluster cannot drift one keyword at a time from “crm” to “crm jobs”.
• Strictness is how many of the top 10 two keywords must share: 3 by default, from 2 (looser, bigger clusters) to 5 (stricter, smaller clusters).
• A keyword that fits no cluster is compared with the 2 or 3 nearest clusters by shared words and results, and an AI model (Jev, from TypeSafe) picks the one it fits, or none. It can only choose between clusters the results made. It never makes a cluster or renames one. Each keyword it places is marked “Placed by AI”.
• Whatever is left goes in Ungrouped. A keyword there may need a page of its own, or no page at all.
• Search intent is read for each cluster’s biggest keyword, with the same labels as the other tools on this site.

Files it reads

• CSV, TSV or TXT, in UTF-8 or UTF-16, with commas, semicolons or tabs. Excel files (.xlsx) too, the first sheet.
• The keyword column is found by its name (Keyword, Top queries, Query, Search term) and the volume column the same way (Volume, Search Volume, Avg. monthly searches), in exports from Ahrefs, Semrush, Search Console and Google Ads Keyword Planner. You can pick other columns before you run it.
• A volume from your file is used as it is. Keywords without one get their volume from DataForSEO for your country and language. Keyword Planner’s ranges, such as “1K – 10K”, are not exact, so those keywords are looked up too.
• Repeated keywords count once, in any capitals or spacing.

Pricing and privacy

• Free up to 100 keywords. Above that, 1 credit ($1) per 1,000 keywords, rounded up, up to 10,000 keywords a run. Credits are charged when the run starts, and given back if it fails.
• The same price whether Google’s results are new or were read in the last 30 days.
• Your keywords are kept only with your report, which only you can open, in your reports until you delete your account. They are never shown to anyone else. Google’s results and search volumes are kept for 30 days for everyone, without who asked.