Release notes

Deep Scan no longer carries a stale failure into the next email, and the domain-age check now honors its timeout end to end.

• Deep Scan no longer shows a leftover failure message after you open a different email
• A failed AI model load is now reported as failed instead of appearing available again
• The domain-age check's timeout now covers its full lookup, not just the final request
• Proton Mail's attachment-preview QR check now always matches the email open when it finishes

The domain-registration check now queries the registry that actually serves the domain, and Deep Scan no longer hangs on the first download.

• The registration-date check now asks the domain's own registry directly, not a shared redirector
• One fewer third party in the lookup path: only the registry serving the domain's ending is contacted
• Deep Scan now reports download progress to a scan that starts while the model is already downloading
• Deep Scan no longer spins forever with no result if you click it during the first download
• Domain endings with no published registry - .ru and .рф among them - can't be dated, and are never flagged for being new

Reporting a scam now counts toward your statistics, so the counters match what you actually caught.

• Reporting a sender as a scam now counts that email toward your scam statistics
• The reported email also appears in your recent-scams list
• Counted once per email, so reopening and re-reporting can't inflate your numbers
• Reporting still doesn't change how future emails from that sender are scored

Fresh-domain reputation: a live check of when a link's domain was registered, plus sharper statistics.

• Live lookup of a domain's registration date against the public registry
• Flags links and senders on domains registered within the last 30 days
• Scam counter and recent-scams refinements in the popup
• Sends only the link's bare domain name to the public registry - never your email content

Shortened-link resolution: see where a short link actually goes before you decide anything about it.

• Detects links hidden behind URL shorteners
• Resolves the real destination without following the link or downloading anything
• Feeds the true domain into the existing reputation checks
• Covers common shorteners such as bit.ly, tinyurl and t.co

QR codes in attachments: decode a QR code hidden in an image attachment, an inline image, a PDF or an Office attachment.

• Decodes QR codes found in image attachments
• Decodes QR codes in images embedded directly in the email body
• Handles QR codes inside PDF and Office attachments
• Works across the supported mail providers, including Gmail, Outlook, Yahoo and Zoho

First-time sender: a caution on email from a domain you have never received mail from before.

• Flags email from a sender domain with no history in your inbox
• Aimed at business email compromise and spoofed invoices
• Works from a per-installation history that is never shared
• Complements trusted contacts, which handles the opposite case

Trusted contacts: the extension learns which senders you actually correspond with, and stops over-flagging them.

• Learns which sender domains you correspond with regularly
• Three or more emails from a domain adds a trusted-contact signal
• A trusted contact lowers the risk score by 10 points
• Fewer false positives on senders you already know

A phishing detection model that runs entirely in your browser, giving a second opinion on emails the rules are unsure about.

• AI Deep Scan, a second opinion on emails that look suspicious but are not caught by the automated rules
• A DistilBERT model trained for phishing detection, about 67 MB, downloaded once
• Runs fully offline in your browser using WebAssembly
• Returns its own verdict, a confidence score and written reasoning
• No email content is sent to any server for analysis

More providers: Yahoo Mail, Zoho Mail including its regional domains, and iCloud Mail.

• Yahoo Mail support
• Zoho Mail support, including its regional domains
• iCloud Mail support
• Provider-specific handling for each service's layout and attachments

Anti-phishing codes: record the secret phrase a service uses, and know whether an email is genuinely from them.

• Configure a personal anti-phishing code for each service you care about, such as a bank, an exchange or an email provider
• Detects the labels providers use, including "anti-phishing code", "security phrase" and "your personal code"
• A matching code shows a green confirmation and lowers the email's risk score
• A wrong or missing code on an email that claims to carry one is treated as a critical scam signal

The first public release: heuristic scanning in Gmail and Outlook, with a colour-coded badge on every email you open.

• Heuristic scanning of every email you open in Gmail and Outlook
• 28 independent checks for sender impersonation, deceptive links, urgency language and suspicious attachments
• A colour-coded badge on each email: green for safe, yellow for suspicious, red for likely scam
• A popover explaining every finding, with the exact text that triggered it
• Mark a sender as safe, or report the email as a scam
• Popup statistics showing how many emails have been scanned