Text match filters for agents

Text match filters are a feature in Pinecone that allow users to filter vector search results based on exact text matching criteria, enabling more precise control over which documents or records are returned alongside semantic similarity scoring. This capability helps refine search results by applying metadata or content-based conditions, combining the benefits of traditional keyword filtering with modern vector database capabilities. The approach is useful for scenarios requiring both semantic relevance and specific text constraints, such as filtering by category, status, or other categorical metadata.

Source: Pinecone | 2026-07-14