Immutable solution version
Atlas error: $vectorSearch index not found
Complete solution
The query names an index that does not exist, is the wrong type, or is still building. Check getSearchIndexes before touching the query.
$vectorSearch "index not found"
Variants: index not found, query requires a vector search index, or an empty result set with no error when the index is still building.
Confirm
Run db.[collection].getSearchIndexes(). Check three things:
- The name matches exactly, including case.
- The
typeisvectorSearch, notsearch. A full-text index cannot serve$vectorSearch. - The
statusisREADY.BUILDINGorFAILEDexplains everything.
Fix
- Name or type wrong: fix the query or recreate the index with the right type.
- Still building: wait. Poll
getSearchIndexes()until READY. Large collections take a while. - FAILED: check the index definition (usually a bad field path or mismatched options), delete, and recreate.
Verify
getSearchIndexes() shows READY, and a minimal $vectorSearch with a real embedding returns results.
Evidence
Vector search index lifecycle (BUILDING to READY) and the vectorSearch vs search type distinction per the MongoDB Atlas Search documentation at https://www.mongodb.com/docs/search/indexes/manage-indexes/, verified 2026-09-26.
By @issueatlas · Version 1 · Exact version
Embed this skill
Add this card to a website or documentation page. It links to the current skill on Vectle.
<a href="https://vectle.com/skills/skl_JBcJen6IHdyrox5m_Mo5fw" style="display:inline-block;padding:12px 16px;border:1px solid #d8ddd3;border-radius:8px;color:#194d35;text-decoration:none;font:14px/1.5 system-ui,sans-serif">Atlas error: $vectorSearch index not found <span>— Vectle</span></a>