About InfoBreach

One identifier in, records out.

There is one motion in this product. You hand us a single identifier and we resolve it against every surface that accepts it.

Four surfaces answer.

Each one holds a different kind of record and accepts a different set of identifiers.

• Breached credentials Search leaked credentials by identifier
• Social accounts Fan one username out across 581 platforms
• Compromised device records Full captures from compromised machines
• People search Assemble a person from US public filings

Four things we will not do.

Constraints on what we are, on what we return, and on what we keep.

• We do not say where the data reaches us from.

  Results name the source a record was recovered from. They do not name how that source reaches us, and neither do exports, nor an answer to asking.

• We apply takedowns to results, not just to records.

  An honoured takedown suppresses the row across all four surfaces, and the counts beside it are recomputed to match. Two things it does not reach: the hourly catalogue totals we cache, and monitors, which keep firing. It also cannot reach the original breached source, which is not ours.

• We do not sell what you searched for.

  Nothing you search becomes a record another customer can find, and we sell none of it. What we do keep: every search is written to an audit log with your IP and user agent, with no expiry, produced only when valid legal process compels us.

• We are not a data broker.

  The four indexes are third-party breach and stealer sources and US public filings, reached through their providers at the moment you search, not compiled into a corpus we own. We do not assemble or sell a database of people. A data broker sells the assembled result; this is the search box in front of sources that already exist.