Wren AI is an Apache 2.0-licensed open-source platform taking a semantic-first approach to text-to-SQL. Its modeling layer translates business terminology to database schemas, enabling accurate query generation even when users don’t know exact table or column names. The three-component architecture includes Wren UI (React-based interface), Wren AI Service (Python LLM orchestration), and Wren Engine (Rust-based SQL engine built on Apache DataFusion). Supports PostgreSQL, MySQL, SQL Server, ClickHouse, Trino, DuckDB, BigQuery, and file sources (CSV, Parquet, JSON). Unique “Generative BI” capabilities include text-to-chart visualization, AI-generated reports, and spreadsheet integration. Docker setup takes approximately 3 minutes. Recently added Model Context Protocol (MCP) support for AI assistant integration with Claude and other tools. Page should cover: semantic modeling tutorial, architecture deep-dive, supported data sources, GenBI features demonstration, MCP integration guide, comparison with Vanna.ai, and self-hosting requirements.