Wren Engine MCP Server

Wren Engine is a semantic engine for MCP clients and AI agents that provides context-aware access to enterprise data with business logic understanding, governance, and support for multiple data sources including BigQuery, PostgreSQL, MySQL, Snowflake, and more.

Features

• Support for 11+ data sources, including BigQuery, PostgreSQL, MySQL, Snowflake, S3, and more
• Semantic understanding of data models with business context
• Reliable calculations and aggregations for reporting
• Business term definitions (active customer, net revenue, customer churn)
• User-based permissions and access control
• Integrates into any MCP client or AI workflow
• Compatibility with modern data stacks
• Semantic-oriented approach that enables AI to understand data models
• Ready for governance with roles and access control
• Runs on Apache DataFusion for semantic SQL

Notes

Currently in beta with releases every two weeks. The project consists of 4 main modules: ibis-server (FastAPI web server), wren-core (semantic engine in Rust), wren-core-py (Python bindings), and mcp-server (MCP server implementation). See mcp-server/README.md for MCP setup instructions. Created by Canner and used in the Wren AI GenBI AI agent.