Sharing a log, stack trace, or .env dump in a support ticket or chat is routine — and it's also one of the easiest ways a real secret ends up somewhere it shouldn't. This tool masks the middle of a key while keeping just enough visible (first/last few characters) for someone to confirm "yes, that's the right key" without ever seeing the whole thing.
What Gets Detected Automatically
| Pattern | Example prefix |
| Stripe keys | sk_live_, pk_test_, rk_live_ |
| AWS access key IDs | AKIA... |
| GitHub tokens | ghp_, gho_, ghu_, ghs_, ghr_ |
| Slack tokens | xoxb-, xoxp-, xoxa-, xoxr-, xoxs- |
| Google API keys | AIza... |
| JWTs | eyJ...eyJ...signature (three dot-separated segments) |
| "Sensitive-looking" lines | Any NAME=value or NAME: value where NAME contains KEY, SECRET, TOKEN, PASSWORD, or AUTH |
This list can't cover every possible key format. Always glance over the redacted output before sharing it — treat this as a strong first pass, not a guarantee.
One-Way, By Design
There's no "unmask" button, and there never will be — the removed characters aren't encoded into the output in any recoverable form. That's deliberate: a redaction tool that could be reversed by anyone who sees the masked output wouldn't actually protect anything.
Frequently Asked Questions
Is this reversible — can someone un-redact the masked key?
No. This is one-way masking for display purposes, not encryption. The original characters that get replaced with dots are simply discarded, not encoded anywhere in the output.
Is my key or log text sent anywhere?
No. Everything runs as JavaScript in your browser. Nothing is uploaded, logged, or transmitted.
What does auto-detect actually look for?
Known key formats (Stripe, AWS access keys, GitHub tokens, Slack tokens, Google API keys, JWTs) by their characteristic prefixes, plus any line that looks like NAME=value or NAME: value where NAME contains words like KEY, SECRET, TOKEN, PASSWORD, or AUTH. It deliberately leans toward over-redacting rather than missing something.
Will it redact things that aren't actually secrets?
Sometimes, yes — for example a variable named NOT_A_SECRET_VALUE would get flagged because it contains the word SECRET. That's an intentional trade-off: for a redaction tool, over-redacting a harmless value is a far smaller problem than missing a real one.
Should I use this instead of rotating a leaked key?
No. If a real secret was ever exposed — committed to git, posted publicly, pasted somewhere it shouldn't have been — rotate it. Masking it for a future share doesn't undo an earlier leak.