Keyword Clustering Tool
Paste a keyword list, optionally with search volumes, and group it into clusters of keywords that share their important words. Each cluster gets a suggested page or topic name, a keyword count and a total volume, so you can see which page should target which keywords. Matching handles Turkish letters and common Turkish suffixes. Everything runs in your browser. For the whole research process, see the keyword research guide.
Clusters
Unclustered keywords (3)
These keywords did not match any other keyword at this threshold. They may be a separate intent worth their own page, or the threshold may be too high.
Grouping is based on word similarity, not on search results. Searches that share words but differ in intent (e.g. "what is" and "pricing") may still need separate pages. Your list is processed only in your browser.
How to use it
- Paste one keyword per line. To include volume, add it after a tab, comma or semicolon (e.g. "seo agency, 2400"). Two columns copied from a spreadsheet or an Ahrefs, Semrush or Search Console export work as-is.
- Adjust the similarity threshold. Higher values give smaller, tighter clusters; lower values merge more keywords into broader topics.
- Review each cluster: does it really share one search intent? Split clusters that mix, for example, "what is" questions with service or price searches.
- Download the CSV and turn each cluster into one page in your content plan, using the suggested name as a starting point.
Frequently Asked Questions
What is keyword clustering and why do it?
Keyword clustering groups keywords that can be answered by the same page. Instead of writing a separate page for every variation, you build one strong page per cluster, which avoids pages competing with each other (cannibalisation) and makes the site structure clearer. It is a core step of keyword research.
How does the tool group keywords?
Each keyword is lowercased with Turkish rules (İ to i, I to ı), optionally folded to ASCII (ş to s), cleared of conjunctions and prepositions such as "ve", "için", "the" and "for", and reduced to simple stems by conservatively removing common suffixes (lar, ler, lık, ı, sı, da, den and so on). The highest volume keyword that is not yet in a cluster becomes the head term, and every remaining keyword whose stem overlap (Jaccard similarity) with it reaches the threshold joins its cluster. Nothing is sent to a server.
Which similarity threshold should I choose?
50% is a good start: "seo agency" and "istanbul seo agency" end up together, while "seo" and "seo agency" stay apart. For a landing page plan where each page targets a narrow intent, try 60 to 70%. For broad topic groups or a content hub plan, try 30 to 40%. Results depend on the head term, so review clusters before you use them.
How is this different from SERP based clustering?
This tool compares the words in the keywords. SERP based tools compare the search results: if two keywords show many of the same URLs, Google sees them as one intent. Word based grouping is fast and free, but it cannot tell, for example, that "what is seo" and "seo pricing" have different intents. Check each cluster with the search intent guide in mind and confirm important ones by looking at the actual results.
Where can I get a keyword list with volumes?
Search Console (queries your site already appears for), Google Keyword Planner and paid tools such as Ahrefs and Semrush all export lists you can paste here. For a comparison see the best keyword research tools. Without volumes, the tool still works; clusters are then ordered by keyword count.
How do I turn clusters into a content plan?
Treat each cluster as one page and the suggested name as a working title. Group related clusters under a pillar page and link them to each other; this is how topical authority is built. Start with clusters that have high total volume and a clear intent you can serve well.


