MindsDB Query Engine

MindsDB Query Engine gives developers building AI agents one SQL interface for business data and document search. It runs as a standalone server and can also provide data access to MindsHub agents.

• Queries across live sources: Query connected databases, warehouses, applications and files, including MongoDB, and combine results without first importing the source data into a separate pipeline.
• Document search: Index content in knowledge bases and retrieve it by meaning, keyword or a combination of both, with metadata filters to narrow results.
• Hybrid retrieval: Combine vector similarity with BM25 keyword matching for searches that include natural language alongside exact codes, identifiers or acronyms.
• SQL client access: Use the built-in browser SQL editor or connect through MySQL- and PostgreSQL-compatible clients such as DBeaver, SQLAlchemy and BI tools.
• Saved queries and scheduled work: Organize work into projects, save queries across sources as views, and schedule SQL jobs to refresh knowledge base indexes.

You can self-host the engine with Docker or run it locally as a Python package. Live queries leave the original data in its connected source; document search creates a separate index. Knowledge bases need an embedding model and a vector store, such as pgvector, and can optionally use a reranking model.