Take the personal data out before you share it

The moment this matters is ordinary and frequent: you need to show somebody a real example. A support ticket that demonstrates the bug, a log with the failing request in it, a transcript that proves the point. The example is only useful because it is real, and it is only shareable if the real parts about people are taken out first.

Doing that by hand fails in a specific way. You catch the obvious ones, the email address and the phone number, and you miss the postcode in line four hundred, because the eye stops looking once the page feels clean. Scale makes it worse rather than better, and the miss is the only part anyone will remember.

This applet replaces what it finds with consistent placeholders, and the consistency is the useful part rather than the removal. If the same address becomes the same label everywhere it appears, the shape of the conversation survives: you can still see that the person who complained on Tuesday is the person who was charged twice on Monday. Redaction that destroys that relationship also destroys the reason you wanted the example.

It is honest about its scope, which matters more than coverage. Pattern matching finds things with a shape: addresses, numbers, postcodes, card numbers. It does not find a person's name, because a name has no shape, and it will not know that a reference number is identifying in your particular system. Anything you share still needs a human to read it. This gets you from unreviewable to reviewable, which is the useful step, not from unsafe to safe.

That distinction is worth being firm about. Pseudonymised data is still personal data under UK GDPR when the original can be recovered, and a scrubbed transcript sitting next to the ticket it came from is exactly that. What the applet reduces is the risk of casual exposure when something is pasted into a chat, a ticket or an email, which is where most of the accidental disclosures actually happen.

Running locally is not a nice extra here, it is the entire point. A tool that removes personal data by uploading the personal data has solved nothing. Everything happens inside this page, which is the only arrangement that makes sense for the job.