[ tool ]
Keyword clustering tool
Paste a keyword list, one per line. The tool groups them by word overlap. No login, runs in the browser.
[ input ]
threshold: 0.3 (higher = tighter clusters)
[ output ]
Paste keywords on the left to see clusters here.
[ how this works ]
Word overlap, not embeddings
Each keyword is tokenised. Stopwords and very short tokens are dropped. The tool walks through your list and, for each keyword, checks Jaccard overlap (shared tokens divided by total unique tokens) against every existing cluster. If the best match is above the threshold, the keyword joins that cluster. Otherwise it starts a new one.
Cluster labels come from the most frequent term in the cluster. A higher threshold gives tighter clusters with more groups. A lower threshold gives looser, larger groups.
This is a heuristic, not search intent classification. For tight semantic clustering you want embeddings. For sorting a few hundred keywords into rough buckets, this is enough.