MCP Toolbox for Databases – Connecting AI Agents to Enterprise Data

Stenal P. Jolly, Senior Cloud Architect at Google Cloud, and Anubhav Dhawan, Software Engineer at Google, introduced MCP Toolbox for Databases—an open-source, enterprise-grade server built around the Model Context Protocol (MCP). In this PyData Milton Keynes meetup session, we explore the Model Context Protocol (MCP) and demonstrate how to build robust, scalable toolboxes that securely connect autonomous AI agents directly to enterprise databases (SQL, PostgreSQL, BigQuery, and more). Learn the foundational concepts behind MCP, explore live architectural demos, and discover practical design patterns for querying, managing permissions, and integrating enterprise data stores into your modern LLM pipelines. ⏱️ Timestamps / Chapters 00:00 – Welcome & Introduction to PyData Milton Keynes 03:45 – The Problem: Connecting AI Agents to Enterprise Data Securely 10:15 – What is Model Context Protocol (MCP)? Core Architecture Explained 22:30 – Building the MCP Toolbox: Tools, Resources & Prompts 37:50 – Live Demo: Querying Databases via Autonomous AI Agents 55:10 – Security, Governance & Permission Handling in Production 1:18:25 – Audience Q&A & Discussion 1:45:00 – Community Announcements & Wrap-up (Feel free to adjust the timestamp markers to match your exact talk structure!) 🔗 Resources & Links 🌐 PyData Milton Keynes Meetup: [https://www.meetup.com/pydata-milton-...] 💻 Code & Slides Repository: [coming soon] Lab: https://codelabs.developers.google.co...