MySQL

MySQL backs a meaningful share of the application-tier data in the wild — particularly in legacy operational stacks, WordPress/Drupal ecosystems, and orgs that standardized on the LAMP stack years ago and never had reason to migrate. The integration treats MySQL as a live source for chats, canvases, and functions, with cross-database queries for the multi-tenant deployments where one user has access to several DBs.

What you can do

• Pull live records into a chat — "show me last month's orders by region" returns a structured table from your live MySQL, no separate query tool.
• Run SQL from a function — joins, aggregates, subqueries — anything MySQL can express, including JSON column handling.
• Cross-database queries — for the multi-tenant model where the same user can join across multiple databases.
• Full-text search built into MySQL, exposed without a separate search service.
• Transaction management — multi-statement writes wrap in a transaction; rollback on error is the default.
• Schema introspection — table and column metadata queryable as structured data.

How to connect

Standard MySQL wire protocol over TCP. Provide host, port, username, password, and optional database name. Leave the database field blank if you want cross-database queries.

The integration honors whatever permissions the database user has — grant only what the workflows need.

Connect MySQL in myai

Common use cases

• Legacy-stack reporting — operational MySQL data joined with warehouse data in a single canvas, no per-system reporting tier needed.
• Customer-facing dashboards generated from MySQL without exporting to a BI tool.
• Daily anomaly digests — scheduled function reads the previous day's MySQL state, surfaces unusual patterns, posts to the team's chat.

Reference

• Connection model: MySQL wire protocol (TCP)
• Auth: username + password (TLS strongly recommended for cloud instances)
• Scope: ~8 operations spanning SQL execution, schema introspection, transactions