Keyword Clustering Tool
Group a flat keyword list into themed clusters by their shared significant words — for SEO content planning.
Quick answer: Group a flat keyword list into themed clusters by their shared significant words — for SEO content planning.
Last updated
Frequently asked questions
- What does 'cluster by shared word' actually mean?
- We find the most-frequent significant word across all your keywords and group every keyword that contains it. We repeat on the remaining keywords, and so on, until every keyword belongs to a cluster.
- What's the goal of keyword clustering for SEO?
- One cluster ≈ one page. Keywords in the same cluster compete for the same intent, so writing one well-targeted page that answers all of them outranks ten thin pages each targeting a single phrase.
- Should I keep stop words?
- Usually no — stop words like 'the' and 'for' would over-cluster everything. Toggle 'Ignore stop words' off only if you're working in a language whose function words carry semantic weight.
- Is this the same as semantic clustering using embeddings?
- No — true semantic clustering uses ML embeddings and finds related-meaning groups even when the words differ. This tool is a fast, deterministic, surface-form-based clusterer that works without any model. For most keyword research it's good enough; for nuanced topical mapping, a vector clusterer is better.
- Why are some keywords in single-keyword clusters?
- When a keyword shares no significant token with any other, it falls into its own cluster. Those usually represent unique intent that deserves its own page or are good candidates to drop from your list.
- How many keywords can I cluster?
- Several thousand keywords cluster in well under a second. Beyond that, the rendering of cluster cards becomes the bottleneck, not the algorithm.
- Will it handle non-English keywords?
- Yes — tokenisation uses a Unicode word-boundary regex, so any language with letter-and-digit-based words works. The English stop-word list won't help in other languages; toggle it off.
- Is my keyword list uploaded?
- No — clustering is a pure JavaScript algorithm in your browser. Even competitor keyword lists are safe to paste here.
- Can I export the clusters as CSV?
- Use the 'Copy as text' button — each cluster has a header line and its keywords, easy to paste into a spreadsheet. A direct CSV export is on the roadmap.
- Why is the cluster size sometimes lopsided?
- Greedy clustering grabs the biggest theme first, so the largest cluster often has many keywords while the rest have few. Re-cluster a single large bucket to break it down further.