Introspects live databases to construct a Kuzu-backed schema graph, serving as an MCP server for AI tools.
Sponsored
Pretensor OSS automatically introspects live database schemas from sources like PostgreSQL, Snowflake, and optionally BigQuery. It then constructs a comprehensive knowledge graph using Kuzu, detailing tables, columns, foreign keys, inferred joins, and other crucial metadata. This robust graph is exposed to AI tools via an MCP (Model Context Protocol) server, enabling AI agents to seamlessly query schema context and perform searches without directly issuing raw SQL commands against the underlying graph store. It empowers data analysts, engineers, and architects by providing grounded schema context for AI-driven data exploration and interaction, eliminating the need for manual DDL sharing with LLMs.
Key Features
01Exposes schema context to AI tools through an MCP server
02Generates a Kuzu knowledge graph of database tables, columns, and relationships
03Offers BM25 keyword search over table and entity metadata
04Automatic schema introspection for PostgreSQL, Snowflake, and BigQuery
050 GitHub stars
06Provides tools for schema inspection, Cypher querying, and SQL validation against the indexed graph
Use Cases
01Enabling data analysts to explore data warehouses using AI assistance
02Automating the provision of database schema DDLs for AI tools and chatbots
03Providing grounded and accurate schema context for AI agents used by data architects
