Last updated: 25 July 2026 Extension version: 2.0.0
The extension checks whether the domain you are visiting imitates a known brand. To do that it sends the hostname only to a checking service. It does not send the page path, query string, or page content. Nothing is stored on the server. There are no accounts, no analytics, and no third parties.
When you visit a page, or press Check this page, the extension may send a single piece of information to its checking endpoint:
example-login.com.It does not transmit:
Requests carry no user ID, no device ID, and no session token. Requests from different users are indistinguishable from one another apart from network-level metadata.
The extension ships with a built-in list of widely-used domains (search engines, major cloud and developer platforms, banks, universities, news sites, and similar). Pages on those domains are resolved **entirely inside your browser and their hostnames are never sent anywhere at all**. In ordinary browsing this means most of the sites you visit never leave your device.
You can read the exact list in background.js in the public source repository.
On the server: nothing. The checking service is stateless. It holds no database and writes no records of the hostnames it receives. Its logs record only the verdict (whether a lookalike pattern matched) and which rule matched. Hostnames are deliberately excluded from those logs.
In your browser: the extension caches verdicts for hostnames you have visited, in Chrome's local extension storage, so repeated visits do not need a repeated request. This cache holds at most 300 entries, expires after 30 minutes per entry, never leaves your device, and is deleted when you remove the extension. You can clear it at any time by removing and reinstalling the extension.
tabs — to read the address of the tab you are viewing, so its domain canstorage — to keep the local verdict cache described above.notifications — to show a desktop notification when a domain appears toIt checks domain names for imitation patterns. It does not detect phishing hosted on a legitimate platform under an unrelated subdomain (for example, a credential form hosted at a random subdomain of a free hosting provider), and it does not inspect page content. It is not a replacement for Google Safe Browsing, antivirus software, or your own judgement. Measured performance figures and known limitations are published in EVALUATION.md in the source repository.
The extension and the checking service are open source and can be audited: https://github.com/abhishekjaden/Phish-Sentry
If this policy and the code ever disagree, the code is the authority — please open an issue and the policy will be corrected.
Material changes will be accompanied by a version increment here and in the extension. The extension will not begin collecting a new category of data without this policy being updated first.
Abhishek Jaden — via the GitHub repository above.