Personal runs the same AI engine as Business

MxGuard Personal is not a cut-down filter. It runs the identical hybrid-AI engine used by MxGuard Business — the only differences are the size limits and the add-ons, not the quality of the filtering.

The same two-layer brain: - A LightGBM gradient-boosted classifier scores around 95% of your mail decisively, in a median of 3.4 milliseconds, using roughly 120 features drawn from the envelope, headers, body, URLs, attachments and sender reputation. - The borderline 5% — the carefully written, targeted messages — are handed to Claude, Anthropic's large language model, which reads them for intent and impersonation before a verdict is reached.

The same supporting checks: ClamAV attachment scanning, SPF/DKIM/DMARC enforcement, DNSBL and URIBL threat feeds, and ARC sealing through forwarding.

The same training. The classifier learns from a global honeypot network, thirty years of UK ISP ham and public reference corpora, so it keeps up with new attacks within hours. A personal mailbox gets exactly the same protection a business does.