Query database schemas and generate valid SQL for any RDBMS

SchemaCrawler is a free database schema discovery and comprehension tool that can search database schema objects using regular expressions and output schema and data in a readable text format, aimed at data governance use cases. This project is a bare MCP client wrapper that connects to the separate SchemaCrawler AI MCP Server, letting an AI coding assistant work with SchemaCrawler in "Agent" mode inside Visual Studio Code.

Use it when you want an AI agent running in VS Code's Agent mode to answer questions about a database's structure - which tables or columns match a pattern, what the schema looks like - by pattern-matching schema objects rather than writing SQL by hand. It is scoped specifically to VS Code's Copilot Chat MCP integration; it is not useful in a client that doesn't support MCP, and it doesn't replace direct use of the SchemaCrawler CLI for users who don't need an AI agent in the loop - that path, along with the rest of this setup, is documented in docs/getting-started.md.

The underlying SchemaCrawler AI MCP Server is distributed as an official Docker-verified image on the Docker MCP Catalog (mcp/schemacrawler-ai), with early-release builds published separately under schemacrawler/schemacrawler-ai. SchemaCrawler's core capability - searching schema objects with regular expressions and producing readable text output of schema and data - is exposed to the AI agent through this MCP integration.

Install Docker, Docker Compose, and Visual Studio Code, then read Microsoft's "Use MCP servers in VS Code" documentation before cloning this project and opening it in VS Code. Follow the project's own getting-started documentation in its docs/ folder for the remaining setup and Agent-mode configuration steps.

Developers and data engineers who want an AI coding agent in VS Code to explore and understand a database's schema - searching for tables, columns, or other objects by pattern and reviewing readable schema output - as part of a data governance or database-comprehension workflow.