Paste an email. See what it’s really asking you to do. It’s checked on your device.
A Scan email button on every message you open.
Download the add-in How to installOpen it, or add it to your home screen. Works offline after the first visit.
Open the web appThe whole scanner in a single HTML file (about 4.6 MB). The online checks (QR pictures, known phishing sites) need the web app.
Download the fileFor Outlook on the web and Outlook for Windows and Mac, with a Microsoft 365 or Outlook.com account. Some work accounts block custom add-ins; your IT admin can install it for everyone from the Microsoft 365 admin centre. Email content never leaves Outlook. Known-phishing-site checks request only a hash-bucket number; the optional domain age check sends domain names to tripass.com.au.
Sign in, pay, change bank details, call a number, scan a QR code, buy gift cards, keep it quiet. In 18 languages.
The sender’s real website is fine. A look-alike, a known phishing site or a free hosting page is not.
Secrecy, “I can’t take calls”, pressure. “Urgent” alone is never enough.
It also reads QR codes in pictures, checks attached web pages for fake sign-in forms, and checks links against about 730,000 known phishing sites without revealing which site it’s checking.
| Test | Before | Now |
|---|---|---|
| 816 real phishing emails, 2020–2023, never seen in training | 67.8% flagged | 81.6% |
| 813 real phishing emails, 2024–2025 | 60.5% | 79.1% |
| 915 spear-phishing emails (synthetic) | 43.0% | 96.6% |
| 2,223 Turkish phishing emails (synthetic) | 3.0% | 86.6% |
| 4,150 real genuine emails (SpamAssassin, 2002) | 1.4% wrongly flagged | 1.7–2.1% |
| 1,500 real genuine emails (Enron, 2001) | 0.5% | 0.5–0.8% |
| 20,000 real genuine links | 12.2% wrongly flagged | 1.8% |
| 120 hand-written business emails | 0 false alarms, all phishing caught | Same |
Measured 5 October 2026.
Training data: Jose Nazario phishing corpus (CC BY 4.0); FERB, GECAD/Polytechnic of Porto (CC BY 4.0); OLTA-TR (CC BY 4.0); Phishing Validation Emails, Miltchev et al. (CC BY 4.0); PhiUSIIL, Prasad & Chandra (CC BY 4.0); Majestic Million (CC BY 3.0); Phishing.Database (MIT); marketeam Marketing-Emails (MIT); free-programming-books (CC BY 4.0); public-apis (MIT); Wikimedia; 18F, Digital.gov and USDS (public domain); GOV.UK Design System (MIT); Project Gutenberg (public domain). QR reading: jsQR (Apache 2.0).