Self-hosted OSINT exposure checker — is your email, username, phone or password leaked? Keyless sources: breaches, infostealer logs, COMB credentials, HIBP k-anonymity, 200-platform username grid. Docker one-command install, no API keys, no AI, Apache-2.0.
FAQ
Does my password ever leave my machine?
No. Your browser hashes it with SHA-1 locally and sends only the first 5 hex characters of that hash. The server fetches the hash-suffix range for that prefix from HIBP's free Pwned Passwords API and returns it; your browser does the final match locally. The full hash — and the password itself — never reach the server. This is the same k-anonymity design used by 1Password, Bitwarden and haveibeenpwned.com's password check.
Can I check someone else's email/phone/username?
The tool is for checking identifiers you are authorized to assess — yours, or your organization's with permission. The README and this FAQ say it plainly: using exposure data to harass, dox or attack people is abuse, and the tool gives you aggregate counts and public-record signals, not ammunition.
Why is domain/website scanning not included?
Scope discipline. Domain/website vulnerability scanning is a different job with a different engine — ours ships as the Website Vulnerability Scanner Community Edition (same license philosophy). The two repos cross-link each other.
Is this an AI tool?
No. Zero AI, zero ML, zero external "smart" services. Every check is a deterministic, reproducible HTTP lookup with honest three-state verdicts (exposed / clear / unavailable). If a source is down, the card says so instead of guessing. That is a feature.
Why do some cards say "unavailable"?
Honesty over theater. Public APIs rate-limit, change, or hiccup. A source that could not
answer is shown as unavailable — it does NOT count as "clear", and it does not kill the
rest of the check (per-source isolation). Re-run the check to retry.
Where does the data come from?
Keyless public sources — breach databases (XposedOrNot, HIBP Pwned Passwords), infostealer corpus (HudsonRock), the COMB credential corpus (ProxyNova), disposable-email databases, the Maigret/WhatsMyName 200-site username corpus, Bluesky, Hacker News, GitHub public search, Shodan InternetDB, ipapi.is, DNS blocklists (Spamhaus ZEN, SpamCop, Barracuda) and more. Full table with attribution: SOURCES.md.
Can I turn sources off?
Yes — Settings page → per-source toggles. Disabled sources are never contacted.
Is it really self-hosted with no API keys?
Yes. Every source in the Community Edition is keyless. Clone, ./install.sh, sign in,
check. There are no license gates and no scheduled phone-home (v2): the optional free
community key only powers update notifications, and the license client only talks to the
server when you explicitly activate/refresh/deactivate a key (see PRIVACY.md)
— the code is short and auditable.
How is this different from holehe / mosint / sherlock?
Those are good tools (some now unmaintained) that each cover one slice — holehe: email→registered-website accounts; sherlock/maigret: username→platforms. Exposure-Check-OSINT covers email and username and phone and password and IP in one self-hosted dashboard with an honest exposure score, history, printable reports and per-source verdicts. It is a maintained, complete product rather than a CLI one-trick pass.
What license is this?
Apache-2.0 (application code). The username site corpus carries upstream attribution
(maigret MIT, WhatsMyName CC BY-SA 4.0) — see storage/app/maigret/sites.json header.
Full details in the LICENSE file and SOURCES.md.