Keyword Clusters

Clusters group semantically similar keywords into topics using vector embeddings. Each cluster has aggregate metrics and an AI-generated label (Ultra plan).

Prerequisites

• Read thekeyword-setup first for auth and configuration
• Requires Pro+ plan (Pro, Ultra, Enterprise)

How It Works

1. Analyzed keywords are automatically embedded as vectors
2. Each analyzed keyword becomes a cluster seed
3. Similar keywords (≥80% cosine similarity) from the full database are pulled in
4. Small clusters merge; large ones split
5. Ultra users get AI-generated 2-4 word topic labels

Clusters generate automatically after keyword analysis when 10+ un-clustered keywords accumulate.