Run knowledge search on OpenSearch, with hybrid built in

OpenSearch is now a supported vector database in Agno through agno.vectordb.opensearch. It handles vector, keyword, and hybrid search, in both sync and async variants, so you can back a knowledge base with OpenSearch whether you want semantic retrieval, plain lexical matching, or the two combined.

from agno.knowledge.knowledge import Knowledge
from agno.vectordb.opensearch import OpenSearch
from agno.vectordb.search import SearchType

vector_db = OpenSearch(
    index_name="agno_docs",
    search_type=SearchType.hybrid,
)

knowledge = Knowledge(vector_db=vector_db)

‍

If OpenSearch already sits in your stack for search or log analytics, this lets you point an agent's knowledge at it instead of standing up a separate vector store to run alongside it. Install it with the agno[opensearch] extra, and the included run_opensearch.sh gets a local instance going for testing.

‍

See the OpenSearch db cookbook for reference.