Context-aware detection
Warning signs are weighed together with what your inbox shows, such as whether you've heard from this sender before.
Chrome extension for Gmail
Your inbox.
Your defense.
Spot suspicious emails before they trick you. PhishGuard analyzes phishing signals, suspicious links, sender identity, and inbox context to help you make safer decisions in Gmail.
- Works locally by default
- AI review only if you turn it on
- Every score comes with reasons
Explainable warnings
Every score comes with the specific signals behind it and a practical next step, not just a red light.
Optional AI analysis
Turn on Gemini review for a second opinion on emails the local checks find uncertain. It's off until you enable it.
Features
Built to explain, not just to alarm
PhishGuard runs inside Gmail and shows its reasoning, so you can decide with the evidence in front of you.
Phishing signal detection
Flags urgent or threatening language, requests for passwords, payment or verification codes, link shorteners, punycode, raw IP addresses, brand-lookalike domains, and links that don't match the sender.
Inbox context
Compares the open email with the recent messages listed in your Gmail inbox: a brand claimed from the wrong domain, a sender you've never heard from, or a delivery or invoice with no matching order.
Explainable risk scores
A compact badge on the email expands into a 0–100 score, a plain-language verdict, the signals grouped by category, and a recommended action.
Optional AI review
When you opt in, emails with an uncertain local score get a Gemini review through a PhishGuard server you run. Clear-cut cases are decided locally.
Privacy-conscious design
The Gemini API key stays on the server, never in the extension. Email content is only sent for AI review after you turn it on, and your other inbox emails are never sent.
Graceful fallback
If AI review is off, not needed, times out or fails, you still get the full local result, and the panel says which one happened.
How it works
Three steps, inside the inbox you already use
-
1
Open an email in Gmail
PhishGuard checks the open message automatically with its local rules and your inbox context. A small badge appears next to the subject.
-
2
Review the score and signals
Select the badge to see the risk score, the verdict and each warning sign it found, grouped so you can scan them quickly.
-
3
Decide: trust, verify or avoid
Follow the recommended action. When in doubt, contact the sender through a channel you already trust, not the links in the email.
Security & privacy
Clear about what stays and what's sent
Here's exactly how PhishGuard handles your email, based on how the extension is built today.
Local by default
Rules and inbox-context checks run in your browser. With AI review off, no email content leaves your device.
AI review is opt-in
When enabled and needed, the email's subject, sender, body (up to 12,000 characters), up to 50 links, and PhishGuard's short findings go to your PhishGuard server, which forwards them to Google Gemini. Your other inbox emails are never sent.
The API key stays on the server
The Gemini key lives in the server's configuration. The extension never sees it and only talks to the local server you run.
Estimates, not guarantees
A score is a heuristic estimate. A low score doesn't prove an email is safe, and a high one doesn't prove it's malicious. Always verify unexpected requests independently.
Get PhishGuard
Install PhishGuard
Chrome Web Store
Coming soonPhishGuard isn't published on the Chrome Web Store yet. One-click install will be available here once it is.
Add to ChromeDeveloper preview
Available nowLoad the extension from its source code in Chrome's developer mode:
- Download or clone the source code from GitHub.
- Open
chrome://extensionsand turn on Developer mode. - Select Load unpacked and choose the project folder (the one with
manifest.json). - Open or refresh Gmail and open an email.
Optional AI review also needs the PhishGuard server running on your computer with your own Gemini API key; see the README.
View source on GitHubFAQ
Questions, answered plainly
What is PhishGuard?
A Chrome extension that checks the email you open in Gmail for common phishing signs and shows a risk score with the reasons behind it. It combines local detection rules, context from your inbox and, if you choose, an AI review by Google Gemini.
Does it work with Gmail?
Yes. It currently works in Gmail on the web (mail.google.com) in Chrome. Other email providers and the Gmail mobile apps aren't supported.
Does it use AI on every email?
No. AI review is off by default. If you turn it on, PhishGuard only asks Gemini about emails whose local score is uncertain (20–59). Emails the local checks score as clearly low or clearly high risk aren't sent.
What email data is shared when AI analysis is enabled?
For an email that needs AI review: its subject, sender, body (up to 12,000 characters), up to 50 links, and PhishGuard's own short findings about it. These go to the PhishGuard server running on your computer, which forwards them to Google Gemini under Google's API terms. Your other inbox emails are never sent. Turning AI review off stops further requests.
Can it guarantee that an email is safe?
No. PhishGuard gives an estimate based on the signals it can see. It can miss phishing emails and can flag legitimate ones. Treat its score as one input, and verify unexpected requests for money, passwords or personal details through a channel you already trust.
How do I install it?
PhishGuard isn't on the Chrome Web Store yet. For now you can load it as a developer preview: get the source from GitHub, open chrome://extensions, turn on Developer mode, and choose Load unpacked. See Install PhishGuard for the steps.
Read the warning signs before you click.
PhishGuard puts an explained risk score next to every email you open in Gmail.