A MySQL client for production that runs queries in read-only transactions

A MySQL GUI that is safe to point at production

Most MySQL clients are built around editing: a results grid you can type into, a schema designer, a one-click “drop table”. That is what you want on your laptop database and the opposite of what you want when you are debugging production at 2 a.m.

DevOpsPad opens every MySQL query in a server-side read-only transaction. The editor header says Read-only session, and MySQL itself refuses data changes inside it, so a stray UPDATE without a WHERE clause fails instead of committing. Queries time out after 20 seconds, and the result preview stops at 500 rows or 8 MiB, whichever comes first.

Browse structure before you query

The Explorer lists databases and tables for the connection. Pick a table to open its structure: columns with types, nullability and defaults, the index list, and the DDL returned by SHOW CREATE TABLE. Each query tab keeps its own database scope, so you can keep orders in one tab and billing in another without reconnecting.

SELECT id, customer, total, status, created_at
FROM orders
ORDER BY created_at DESC
LIMIT 100;

Press Cmd+Enter (or Ctrl+Enter) to run. The results grid can be sorted, filtered and paged, a row opens in an inspector on double-click, and Export CSV saves what you see.

Private networks and credentials

If MySQL only listens on a private network, turn on an SSH tunnel. DevOpsPad dials MySQL through the SSH connection itself, with host-key verification against known_hosts or a pinned SHA256: fingerprint. The database password, the SSH password and the key passphrase are each stored in the macOS keychain, never in the profile file.

Pair it with a read-only account

The read-only transaction catches the most common mistakes in the editor, but it is not a complete boundary: MySQL commits implicitly before DDL such as DROP, ALTER or TRUNCATE, so those statements are not reliably blocked by it. The real protection is a SELECT-only MySQL user. Use one for any shared or production database.

Other engines work the same way: see the ClickHouse connector for analytics tables, or the full connector list.