Best practices 10 min read 2,110 words By LaiUse Editorial Team Reviewed 28 October 2026

How to Check Whether a Password Has Been Leaked, and What to Do About It

How to Check Whether a Password Has Been Leaked, and What to Do About It: A password can be strong and still be unsafe.

Why this matters

A password can be strong and still be unsafe. Length, character variety, and resistance to obvious guessing attacks describe how difficult a password is to guess. They do not establish whether the same password has already appeared in a data breach, been stolen from another service, or been exposed in a leaked log. Once a password has been used by a service that was compromised, changing a few characters or adding a symbol does not make that particular password trustworthy again.

The correct response to a suspected leak starts with controlling where the password is entered. Do not paste a real password into an unfamiliar website merely to ask whether it has been breached. A genuine breach check requires comparison against breach data, and a website that asks for the complete password may collect it even if the site promises to protect or discard it. The local tools documented here do not establish a remote breach-database lookup. The password-strength tool is a browser-local, live strength check with a published 12-character limit. The password-generator tool creates passwords locally with a secure random source. The privacy-leak-test tool is also browser-local and uses local storage, with no backend calls documented. That architecture may be useful for local investigation, but it does not by itself prove that a credential appears in an external breach corpus.

The process

  1. Confirm the source of the warning. Begin with the notice from the affected service, a security advisory, or a breach notification. The useful question is not merely “Is this password weak?” but “Was this exact credential associated with a compromised account or system?” A weak password deserves replacement, but weakness and confirmed exposure are different findings. If the warning is vague, review the service’s official account, security, or incident communication rather than sending the password to a third-party lookup page.
  2. Do not test the real password online. Avoid entering a production password into any breach-checking site whose internal process you have not independently established. The tools for the tools here document no backend calls, which means the documented tools do not remotely submit passwords. That does not make arbitrary third-party sites safe. For local diagnostics, use a made-up test password rather than the credential under investigation.
  3. Use the password-strength tool for what it actually measures. Open the local strength check and type a non-sensitive example into the field labeled Password. It updates live and performs its analysis locally. Use the result to understand the character of a candidate, not to infer that a password is “safe.” A high score cannot reveal prior use in a breach, and a low score does not prove that the password has leaked. The published 12-character limit belongs to this tool and should not be treated as a general security threshold.
  4. Identify every place the password was reused. If the compromised password was also used for email, a password manager, banking, work systems, or another service, those accounts need attention as well. Changing only the service named in the alert can leave an attacker with a working credential elsewhere. Prioritize accounts that can reset other accounts, recover email, authorize payments, or administer shared systems.
  5. Change the password at the affected service through its trusted interface. Use the service’s official application, website, or another interface you already know is legitimate. If the service provides a way to end active sessions after a password change, use it. A new password should be unique to that service rather than a small variation of the exposed one, because attackers often try previously observed passwords, substitutions, appended characters, and related patterns against the same account.
  6. Generate a replacement with the password-generator tool. Set Length: characters to an appropriate value and generate the replacement locally. The documented implementation uses a secure random source and has a copy function. These properties make it suitable for creating a new password without sending the value to a server. The tool does not publish a maximum length, so the length value should be chosen based on the service’s requirements and normal password-management limits rather than on an assumed unlimited range.
  7. Store and use the new password uniquely. Transfer the generated value through the page’s copy function, then save it in the appropriate password manager or secure credential workflow. Do not reuse it on another site. If a generated password is longer than the target service permits, generate another one at a supported length rather than shortening it by hand, because manual truncation can discard the randomness that made the password useful.
  8. Revoke access and strengthen account recovery. Review signed-in devices, active sessions, recovery email addresses, phone numbers, application passwords, and delegated access. Remove sessions or applications that you do not recognize. Where the service supports it, enable a second authentication factor or a security key. These actions address the possibility that an attacker retained access after recovering the password.
  9. Watch for follow-on abuse. After changing credentials, review recent sign-ins, password-reset events, recovery changes, financial activity, and messages sent from the account. A breach can enable more than password reuse: attackers may impersonate the account, answer security questions, abuse recovery channels, or target people who trust messages from the compromised account. Treat unexpected requests for money, credentials, or confidential information as suspicious even if they appear to come from a familiar contact.

Worked example

Suppose an online store sends a message stating that an account associated with an old email address was included in a breach. The correct first action is to open the store through its known official route and change the account password there. The reader does not paste the exposed password into a web-based leak checker. Instead, they may enter a made-up value such as correct-horse-battery-staple into the local Password field to see how the strength tool responds. Even if the result looks favorable, the exposed credential remains compromised. The reader also searches their password manager for reuse of that credential and finds that it was used for a personal email account.

They change the password at the store first, then change it at the email provider, because email can be used to recover other accounts. For each new credential, they open the password-generator, set Length: characters to a value accepted by the service, generate a password, and use the available copy function to place it in the password manager. They sign out of other devices, review active sessions, verify that recovery addresses and phone numbers are correct, and enable a second authentication factor where available. They do not treat the local strength display or the privacy-leak-test page as evidence that the original password was absent from a breach dataset; neither documented tool establishes such a lookup.

Over the following days, they check the store and email accounts for unfamiliar sign-ins, recovery changes, and suspicious messages. If the old password had also appeared in a pasted-text collection circulating online, a password change at the breached store would not help accounts that never used it. That is why reuse inventory and account recovery review are part of the response rather than optional extras.

Common mistakes

Limitations of this approach

Frequently asked questions

Can a local password-strength check tell me whether a password was leaked? No. It can assess a candidate according to the tool’s local criteria, but a strong password may still have appeared in a breach, and a weak password may not have. Exposure and guessability are separate questions.

Is it safe to paste a suspected password into a website to check it? Do not do that unless you have independently established how the site handles the value. The mere act of submitting a password can expose it. The browser-local tools documented here are preferable for non-sensitive testing because their tools record no backend calls.

Does the privacy-leak-test page confirm that a password appears in a breach database? The tool does not establish that behavior. It records local processing, local storage, and no backend calls, so its result must not be presented as confirmation of a remote breach match.

What should I do first after receiving a breach warning? Change the affected credential through the service’s trusted interface, then check for reuse on other important accounts. Review sessions and recovery settings, enable stronger authentication where available, and watch for suspicious activity.

Is a complicated variant of my old password sufficient? No. Attackers may try predictable mutations, and the old credential may already be known. Generate an unrelated password with the password-generator tool, whose implementation is documented as using a secure random source.

Why does the generator expose a “Length: characters” setting? It lets the reader choose the size of the generated credential while keeping the output locally generated. Because no maximum is published in the tool, use a length the destination service accepts rather than assuming unlimited input is supported.

Do I need to change an exposed password if I have already enabled two-factor authentication? You should still change it. A second factor may block some automated sign-ins, but it does not make a known password safe, eliminate session theft, or prevent phishing and recovery-channel abuse. Use both a unique password and the strongest practical second factor.

How long should I monitor after changing the password? There is no fixed period established by these tools. Review recent account activity and recovery events, keep notifications enabled, and continue monitoring as long as the affected service or your own risk assessment warrants. The important point is that credential replacement begins the response; it does not end it.

The tools in this guide: password strength · password generator · privacy leak test — each runs in your browser, so your data stays on your device.

Browse every guide · Read the blog · About LaiUse