Open-Source Tools for Postgres & AI

You don't need a specialised "AI Database" to build world-class agentic workflows. In this session, we explore pgEdge's open-source Agentic AI Toolkit - a suite of components designed to connect Postgres with AI agents. The centrepiece is a Model Context Protocol (MCP) server that works with any Postgres instance, whether it's running on RDS, Azure, a local Docker container, or at the edge.

We'll walk through the tools the MCP server exposes to LLMs: schema discovery, safe read-only query execution, query performance analysis, and semantic search via pgvector. You'll see a live demo bridging the gap between existing relational data and AI agents - proving that the future of AI-driven data is open-source, portable, and platform-independent.

Getting started is simple: download the binary from the pgedge-postgres-mcp GitHub repository, add it to your AI client's MCP configuration with your database credentials, and you're ready to go. This works with Claude Desktop, Cursor, Claude Code, Ollama, and any other MCP-compatible client. The toolkit also includes its own built-in natural language agents for the web and CLI, so you don't even need a third-party tool to get started.

As always, we'll have short intros, a presentation, and time for open discussion at the end.