There is a moment in every incident where you are about to paste a log somewhere: a ticket, a chat channel, a support forum, a shared dashboard. And in that log, somewhere in the middle of thousands of routine lines, sits a real secret — an OAuth bearer token, a database password in a connection string, a customer email address, an internal IP. One paste is all it takes to leak it. Sensitive Data Masker exists to make that moment safe.
Paste your log, pick which patterns to apply (IPv4/IPv6 addresses, emails, bearer tokens, API keys, passwords, JWTs, AWS keys, credit-card-like numbers, or your own regular expressions), and run. Every match is replaced by a mask, so the shape of each field survives — you can still see that there is a token on the line and roughly where — but the secret itself is gone. Enable Partial keep prefix and the first few characters of each value remain visible for quick manual triage, which is often the sweet spot for debugging while still protecting the bulk of the secret.
Each of the eight built-in categories runs independently and can be toggled off if it does not apply to your data, and the summary cards report exactly how many of each type were found and masked so you know the pattern actually did something rather than silently matching nothing. The IPv4 and IPv6 address patterns are tuned to avoid firing on plain HH:MM:SS clock readings that happen to share a colon-separated shape with a shortened IPv6 address, so turning IP masking on will not chew through every timestamp in a bracketed or space-separated log format. Custom regular expressions, one per line, extend coverage to anything the built-ins do not anticipate — internal request IDs, vendor-specific key formats, or a connection string shape unique to your stack.
Masking is deliberately local and deterministic: the same input with the same options produces the same output, every time. Nothing is uploaded, nothing is guessed, and nothing is removed except the exact values the patterns identify. The result is a log that is safe to share and still legible to the humans and systems that need it.