BreachDetect
← All posts

9 min read

How to check if your customers' passwords have leaked

A search of your email domain won't find your customers. How to find leaked logins for your website, test them against your accounts and reset the right ones.

By BreachDetect Research

On this page
  1. Why your website is in the leak data
  2. Four ways to check
  3. What to do with the results

Most leak checks start from an email domain. Type in yourcompany.com and you get every leaked login for an @yourcompany.com address. That finds your staff. It will never find your customers, because they sign up with Gmail, Outlook and iCloud addresses. When we looked at leaked logins for 40 big consumer websites, 63% of the email addresses were Gmail.

To find your customers you have to search from the other direction, by the website the login was for. This guide covers how to do that, what each option will and won't tell you, and what to do with the results.

Why your website is in the leak data

When infostealer malware infects a computer, it copies every password saved in the browser. The browser stores each one with the address of the page it was saved on, so the stolen record says exactly which site the login belongs to. Have I Been Pwned describes a typical line like this:

https://netflix.com/login:jane@gmail.com:P@ssw0rd

That's the address, the login and the password. Combolists, the lists credential stuffing tools run through, are increasingly made of lines in the same format, cut down from stealer logs. So if your customers have saved their logins for your site on an infected computer, those logins sit in these lists under your site's address. If you want the background on where the logs come from, start with what stealer logs are.

Companies notice the effect before they understand the cause. When Troy Hunt first loaded stealer logs into Have I Been Pwned in January 2025, he wrote about companies that had seen a rise in account takeovers "and the attackers were getting the password right first go every time!" That's what it looks like when someone is working from a list made for your site, rather than guessing.

Four ways to check

1. Check passwords as customers type them

Have I Been Pwned's Pwned Passwords service holds well over a billion passwords from breaches and stealer logs. Your server sends the first five characters of a password's SHA-1 hash and gets back every hash that starts the same way, along with how often each has been seen. You check for a match on your side, so the password never leaves your server. It's free, needs no API key, and handles more than 18 billion requests a month.

Run it when someone signs up, changes their password or resets it. NIST's password guidance, SP 800-63B, requires exactly this. A site "SHALL compare the prospective secret against a blocklist that contains known commonly used, expected, or compromised passwords." You can also run it at each sign-in and send customers with a leaked password through a reset.

What it can't tell you is whether this customer's login for your site has leaked. It only knows the password has turned up somewhere. A common password fails whoever picks it, and an unusual one copied off a customer's laptop last week passes until it reaches the public lists. And it only works when the customer types their password, so it says nothing about the customers who haven't signed in for months.

2. Your sign-in provider's breached password feature

If you use a hosted identity provider for customer accounts, it may already have something.

  • Auth0 has Breached Password Detection, which compares logins against breaches that have been made public. Auth0's own documentation says detection typically takes "Up to 7-13 months" after a breach. Its Credential Guard add-on for Enterprise customers cuts that to 12 to 36 hours.
  • Okta checks passwords against breached credentials at sign-in and during self-service resets. Its documentation is clear that the feature "doesn't retroactively check for breached credentials."
  • Microsoft Entra ID Protection has a leaked credentials detection, but it's aimed at staff accounts rather than customers.

These are worth turning on. Like Pwned Passwords, they mostly act at the moment a customer signs in, and they work from public breach data rather than searching for your site.

3. Have I Been Pwned's stealer log search by website

In February 2025 Have I Been Pwned added a search that does start from your website. Once you've verified you own a domain, you can ask for the email addresses of everyone whose login for that site turned up in a stealer log it has loaded. Its documentation describes it as a search "performed by a website operator looking to identify which of their customers is likely to be the subject of an account takeover attack."

It comes with the Pro plans, which start at $379 a month billed yearly. It never returns passwords.

That last point matters more than it sounds. With only an email address, you know a customer's computer was infected at some point. You don't know whether the password it gave up is still the one on their account. So you either reset everyone on the list, including people who changed their password long ago, or you reset nobody and send a warning. When Google studied its own leaked-password warnings in 2019, people changed only 26% of the passwords it flagged.

4. A leaked-login monitoring service

The last option is a service that searches stealer logs and combolists by your website and gives you each login with its password, so you can test it against your accounts. That's what we built Customer Protection for. It finds every login saved for your domain and its subdomains, with the page it was saved on and when it was posted. It checks every day and alerts you when new ones appear.

You can see how many there are for your site with the free account takeover check before paying for anything. The full list starts at $82 a month billed yearly, or $99 billed monthly, with a data processing agreement included. See pricing.

Finds logins leaked for your siteGives you the password to testPrice
Pwned PasswordsOnly when a customer types itChecks the password itselfFree
Identity provider featuresPublic breaches, mostly at sign-inNoVaries
Have I Been Pwned website searchYes, from stealer logsNoFrom $379 a month
BreachDetect Customer ProtectionYes, stealer logs and combolistsYesFrom $82 a month billed yearly

Most sites should use the first option whatever else they do. It's free, and it stops new customers starting out with a password that's already in every attacker's list.

What to do with the results

Sort out what isn't a customer

Some of what turns up under your website won't be customer accounts. Logins that use your company's own email domain, or that were saved on staff pages like your single sign-on or webmail, belong to employees and contractors. Take those out and deal with them first, because one of them can open far more than a single account. We wrote up what to do when a staff login is in a stealer log.

Then match the rest against your customer table. Some logins won't match any account. At airlines and hotel groups, most leaked logins are membership numbers rather than email addresses, so match on whatever customers actually sign in with.

Test each password against the account

You don't need to log in with a leaked password to know whether it still works. Your site already checks passwords at every sign-in by hashing what the customer typed and comparing it with the stored hash. Run each leaked password through the same function on your server. It creates no sign-in, trips no lockouts and doesn't show up in the customer's account history.

There are three possible answers:

ResultWhat it meansWhat to do
No accountThe customer never signed up, or used a different loginNothing
Password doesn't matchThey've changed it since it leakedUsually nothing. For stealer log entries, consider telling them their device may be infected
Password matchesAnyone with the list can sign in as themReset now

Delete the leaked passwords once you've done this. You only need the answer.

Reset the ones that match, and sign them out

For every match, force a password change at the next sign-in. NIST's guidance says a site "SHALL force a change if there is evidence that the authenticator has been compromised," and a matching password in a leaked list is that evidence. Big platforms have done this for years. Google said in 2019 that resetting the passwords of accounts caught in other companies' breaches had protected more than 110 million users over two years. Facebook said the same year that it checks whether stolen email and password combinations "match the same credentials being used on Facebook."

Resetting the password isn't enough on its own. Also end every active session, revoke remembered devices and app tokens, and clear any "keep me signed in" cookies. The UK's National Cyber Security Centre gives the same order for any hacked account. Change the password, then make sure every device and app still signed in is logged out.

Treat stealer log entries differently

A login from a stealer log means more than a reused password. It means the customer's computer was infected, and two things follow.

First, the malware usually takes session cookies along with passwords. Google's Chrome team says cookie theft "bypasses two-factor authentication and any other login-time reputation checks," and "the stolen cookies continue to work even after the malware is detected and removed." That's why ending sessions matters as much as the reset.

Second, the new password may not stay private. SpyCloud warns that "resetting credentials before wiping the device is not enough," because many infostealers include keyloggers that capture the replacement. Tell these customers their device needs cleaning before they choose a new password.

Tell customers without sounding like a scam

A message about stolen passwords is exactly what phishing looks like, so write it the opposite way. The NCSC's advice for businesses that text or call customers carries over well to email. Leave out links where you can, and "avoid language that induces panic or implies urgency." Something like this works:

We've reset your password

Your email address and password for [Shop] appeared in a list of logins stolen from other websites or from infected computers. [Shop]'s own systems weren't breached. To protect your account we've reset your password and signed you out on every device.

Next time you visit [Shop], choose "Forgot password" on the sign-in page to pick a new one. If you've used the same password anywhere else, change it there too. If you think your computer may have malware, run a security scan before choosing a new password.

We'll never ask for your password by email or phone.

No link, no deadline, nothing to download. Customers who want to act can go to your site the way they always do.

Check again

New stealer logs and combolists are posted every day. For the 40 big consumer sites in our study, the median site had 1,380 new customer logins posted in a single month. A list you cleaned up in March is out of date by April. Run the check at least weekly, or set up alerts, and work through new matches as they come in.

Check your site for free to see how many of your customers' logins are out there now.

See what has already leaked from your company.

Search your domain across stealer logs and breached databases. Free, no account needed.