Vector Store icon

Vector Store

Store, search and manage vectors on the n8n host filesystem via js-vector-store

Actions12

Search Across

AI-generated

Summary

Searches the specified collections with a query vector and fuses their matches into one ranked list. Provide a vector whose length matches the credential dimension; the operation can return up to the configured limit, ranked using the selected metric.

Inputs

  • Query Vector (required) — Array of finite numbers or a JSON string containing one. Its length must match the vector store credential dimension.
  • Collections (required) — Comma-separated collection names to search. Names are trimmed and duplicates removed; at least one is required.
  • Limit — Maximum number of results, default 50; must be at least 1.
  • Metric — Ranking metric: cosine (default), dot product, Euclidean, or Manhattan.
  • Read Only — Defaults to false. This is a read operation, so enabling read-only mode does not block it.

Output shape

A list of n8n items, one per hit, each containing id, score, and metadata.

Missing metadata is returned as an empty object. Results are fused across collections and do not identify which collection produced each hit; store the origin in document metadata if needed. With Continue On Fail enabled, an error is returned as an item containing an error field.

Links

Discussion