SQL Query Anonymizer

Paste a production query, replace customer PII with stable fake values in this browser, then copy the result into an AI chat or a local EXPLAIN. The SQL shape stays; the people do not.

SQL Query Anonymizer / Data Masker

Replace PII in SQL strings with stable mock values. The query should still parse. Nothing is uploaded.

Original SQL Masked SQL

Heuristic masker, not a compliance product. Read the output before it leaves your machine.

Why logs and queries are different jobs

The Log Anonymizer turns secrets into labels ([EMAIL:a3f2]). That is correct for a Slack paste.

A query you want to run still needs values that look like data: an email that parses, a phone that is still a phone, the same original email mapping to the same fake so a self-join does not fall apart. Labels break EXPLAIN, local seeds, and “will this predicate use the index?”.

This page tokenizes SQL strings and comments, then swaps emails, phones, SSNs, card-shaped numbers, IPs, UUIDs, and name-like literals for mock values. Identifiers (users.email) stay. Heuristics miss odd formats and can over-mask a string that merely looks like a name — read the output.

What it will not promise

It is not a compliance product and not a full SQL parser.
Column names, comments with unusual IDs, and national ID formats it does not know will leak if you do not look.
Closing the tab is the cleanup. Nothing is stored.

For pretty-printing after you mask, use the SQL Formatter.