PostgreSQL Tooling Across AI Editors and Agents | POSETTE: An Event for Postgres 2026

by Matt McFarland

Matt McFarland walks through how PostgreSQL tooling is evolving across AI editors and agent-driven workflows, and how Microsoft is extending capabilities from the VS Code PostgreSQL extension into other environments.

Overview

What problem this talk is addressing

Developers interact with PostgreSQL through a mix of:

As AI-native editors and agent-driven tools become part of everyday workflows, the goal is to make PostgreSQL capabilities available wherever developers (and AI tools) operate.

Foundation: the VS Code PostgreSQL extension

The session builds on the foundation established in the VS Code PostgreSQL extension, and discusses how core database workflows can be surfaced consistently across multiple host environments.

Extending tooling across AI editors and agents

The talk describes extending PostgreSQL capabilities from the VS Code extension to:

The approach uses an MCP server as a shared interface so that multiple hosts can reuse the same underlying PostgreSQL tooling.

Capabilities highlighted

Across these environments, the session calls out PostgreSQL workflows such as:

It also contrasts richer editor environments (VS Code and Cursor) that can support interactive query execution and visualized results, with CLI-based workflows (GitHub Copilot CLI) that can perform similar operations without a visual UI.

Azure integration

The video chapters and abstract indicate Azure Database integration in the VS Code extension, including a demo of Azure server management from VS Code.

Architecture direction

The session closes by discussing how an MCP-backed architecture can help PostgreSQL tooling continue to expand as new AI-driven environments emerge, so core database capabilities can surface naturally wherever AI developers work.

Video chapters