Schema Extractor Agent

This agent retrieves or updates schema information from a connected data source so an LLM can generate accurate SQL. It first checks whether schema metadata is already in ChromaDB; if not, it queries the database, updates ChromaDB, and sends the schema context back for SQL generation.

The agent does this without user prompts, extracting and structuring schema context for downstream agents and tracking changes in the database structure over time. That eliminates the need for manual schema mapping or dependence on DBAs, both common pain points in data querying. It is designed for business users who query data across SQL databases in natural language, which accelerates time-to-insight and broadens data access across the enterprise. It connects to PostgreSQL, MySQL, BigQuery, and SQLite, and integrates with ChromaDB.