VectorXInstall

Summary

Perform a similarity search against a VectorX index.

Inputs

• main (required) — The data to process. Each input item triggers an identical search using the node's parameters.

Output shape

array

For each input item, the node returns one output item per matched vector. Each output item contains the fields id (string), similarity (number), and meta (object).

Examples

Example 1: Find the 5 most similar document vectors to a query embedding, filtering by category.

Set Query Embedding to the query vector (e.g., [0.1,0.2,0.3]), Top K to 5, and Query Filter to [{category: {$eq: 'reference'}}].