//
Adaptive Spam Filter
Overview
We've completely rebuilt how Vern handles spam. The new Adaptive Spam Filter uses two distinct modes — one for strangers, one for people you already trust — so legitimate emails stop getting flagged while real threats still get caught.
✨ What's New
Dual-Mode Detection: New senders get the full spam and phishing analysis. Known contacts switch to compromise detection — checking if their account might have been hacked.
Sender Baseline Learning: Vern studies previous emails from each contact to understand what "normal" looks like for that person — tone, writing style, typical requests.
Anomaly Alerts: If a trusted sender suddenly writes differently, requests unusual actions, or links to suspicious sites, Vern flags it as a potential account takeover.
Consistent Inbox Placement: No more half-in-spam, half-in-inbox chaos. Emails from the same sender land in the same place, every time.
Transparent Reasoning: See exactly why an email was flagged — whether it's a new sender risk or a deviation from a known contact's normal behavior.
⚡ Why It Matters
Traditional spam filters treat every email the same way — which is why your colleague's messages randomly end up in spam while obvious junk slips through. Vern's adaptive approach means trusted senders stay trusted, and you only get alerted when something actually changes.
