Cocoon reads your Search Console data, clusters it by what actually ranks together, and surfaces the new landing pages worth building — without cannibalizing the ones you have.
SERP overlap — not intent — decides what belongs together. Intent is metadata. The expensive SERP checks only run on representative keywords, so a million queries cost a few thousand lookups.
Groups queries that should live on one page, by how much their search results actually overlap. Picks the main keyword and the URL that should own it.
SERP-overlap drivenRuns regex and long-tail patterns on the raw keyword pool to find child-page candidates — jurisdictions, segments, questions, comparisons — your parent page is silently absorbing.
new-page finderBefore recommending a page, it measures overlap with the parent and existing URLs. High overlap? Keep it merged. Clean separation? Build it.
parent-separation scoreEach cluster resolves to one decision — exportable to CSV, XLSX or JSON, with suggested URL, title, H1 and internal-linking anchors.
Real demand, no page that owns it, clean separation from the parent.
A live URL already fits — route the keywords there instead of splitting.
SERP says it's the same page — folding it in beats fragmenting.
A page exists but the parent steals its traffic — a human should decide.