Best on desktop — the editor needs a wider screen. Try ERDone on your laptop.

Pair with Claude

AI design review for the schema in your repo.

A second set of eyes on the design — not the syntax. Claude reviews how your schema is modeled and shows you what to fix before you build on it.

ERDone's Pair with Claude drawer showing AI design-review findings as collapsible cards, each with a Jump to action and a Dismiss button

What it does

Click Pair with Claude and it reviews how your schema is modeled: a table that should be split, a missing index on a column you join on constantly, a relationship pointing the wrong way, naming that won't age well. Each finding is a card you can jump to, act on, or dismiss — dismiss it and it stays gone. It's a review, not a rewrite: you decide what changes.

Not another linter

ERDone already runs deterministic health checks for SQL syntax and dialect correctness. Pair with Claude reviews the modeling underneath — the judgment calls a linter can't make.

Why ERDone, not just the AI

Plenty of tools have bolted on an AI. What they don't have: your diagrams committed to your own GitHub repo with full git history (never on our servers), real baseline-diff migrations that flag every DROP before you run it, and 8 SQL dialects — all for $7/month. Pair with Claude is the newest reason to use ERDone; it isn't the only one.

Who it's for

The engineer who just joined the team and inherited a database design nobody's reviewed in a while.

The lead engineer, pulled in ten directions, who wants a second set of eyes on whether the design will scale.

The engineer who's shipped a migration that went badly because a design problem surfaced too late.

Honest note

Every finding is a suggestion you can dismiss, and it never edits your schema. Your diagrams still live in your repo; running a review sends the schema's DDL to the AI only when you click, nothing stored. On Pro ($7/mo) and Team ($12/mo); free for 14 days on trial.