OpenSearch vs Weaviate API + MCP

OpenSearch

OpenSearch 3.9.0 runs vector, text and hybrid search under Apache-2.0, with an OpenAPI 3.1 description of its REST API and API keys limited by permission and index pattern. The owner runs the cluster, no status page or SLA comes with the software, and no tool in the MCP server sets a read-only or destructive annotation.

Weaviate API + MCP

Built-in MCP server with 4 typed tools, readOnly, destructive and idempotent hints on each, and write tools hidden until enabled. No public status page or incident history.

OpenSearch

1. Create the index with index.knn set to true and a knn_vector field of the right dimension before indexing vectors
2. Put the filter clause inside the knn query for filtering during the search. A filter outside it runs afterwards and can return fewer than k hits
3. Ask a security administrator for an API key and send it as Authorization: ApiKey <token>. The token is shown once and lasts 90 days at most
4. Use size, _source and filter_path to keep responses small, and search_after with a point in time for deep paging
5. Set OPENSEARCH_SETTINGS_ALLOW_WRITE=false on the MCP server unless writes are needed. GenericOpenSearchApiTool can call any endpoint

Weaviate API + MCP

1. Call weaviate-collections-get-config first so the model sees property names before it writes a filter
2. weaviate-query-hybrid defaults alpha to 0.75, so lower it for keyword-heavy queries
3. Turn on the read-only toggle in Weaviate Cloud before giving an agent the MCP endpoint
4. Set object IDs with generate_uuid5 so a retried import doesn't create duplicates
5. Free clusters hold one collection. Use tenants rather than extra collections