MongoDB Atlas Vector Search: index not found is usually a name or status problem

[cristiantumani/corteza.app + luvadlamudi/tamuhack26]: the index name is case-sensitive and must match exactly what the query references. Check the index status in the Atlas UI: Building means queries are not served yet. numDimensions must equal the embedding model's output (1536 for text-embedding-3-small, 3072 for text-embedding-3-large, 1024 for voyage-3-large). Add filter-type fields for metadata paths like workspace_id to enable pre-filtering before the vector search. And check Network Access: a missing IP whitelist entry makes every operation fail or hang with ServerSelectionTimeoutError.

Context: Empty results or 'Cannot find index' errors almost always trace to the index name, its build status, or the dimension config.

Source: Empty results or 'Cannot find index' errors almost always trace to the index name, its build status, or the dimension config.

Verified resolution: [cristiantumani/corteza.app + luvadlamudi/tamuhack26]: the index name is case-sensitive and must match exactly what the query references. Check the index status in the Atlas UI: Building means queries are not served yet. numDimensions must equal the embedding model's output (1536 for text-embedding-3-small, 3072 for text-embedding-3-large, 1024 for voyage-3-large). Add filter-type fields for metadata paths like workspace_id to enable pre-filtering before the vector search. And check Network Access: a missing IP whitelist entry makes every operation fail or hang with ServerSelectionTimeoutError.

Source URL: https://github.com/cristiantumani/corteza.app/blob/HEAD/setup-vector-search.md