Zevr Guard — krystasis

Your browser talks to dozens of companies on every page you visit. Zevr Guard shows you exactly who — on a live world map — and blocks the dangerous ones automatically.

Install once. No setup, no account. Protection starts the moment you add it.

━━━ WHY ZEVR GUARD ━━━

✔ Blocks malware & phishing with a threat list that updates itself daily — no waiting for extension updates
✔ Detects lookalike phishing domains (examp1e.com, exаmple.com, example.com.verify-account.net) with on-device heuristics — caught even before they reach any blocklist
✔ Blocks entire countries with one tap — see a connection you don't like on the map, block everything from that region
✔ Guards password entry: warns before you type a password on an unencrypted page, a suspected lookalike, or a site you've never signed into
✔ Identifies the companies behind 115,000+ tracker signals, not just cryptic domain names
✔ 100% local matching — your browsing URLs never leave your device. No telemetry, ever
✔ Built for Manifest V3 from day one — works today, works after the MV2 shutdown
✔ Fully open source (GPL-3.0) — every claim above is verifiable in the code

━━━ WHAT YOU'LL SEE ━━━

1. Risk score & live tracker list (popup)
   A per-tab risk score, every third-party connection with company and country, and one-tap block/allow controls.
2. Live Globe (side panel)
   Watch your data cross the world in real time, broken down by company and connection count. Screenshot-worthy — literally: a built-in share card turns any page's scan into an image.
3. Auto-block warning screen
   Navigate toward a known malicious domain and Zevr Guard intercepts the request with a clear warning — before the page ever loads.
4. Weekly protection report
   Blocked-per-day chart, top watching companies, most blocked domains. All computed on your device — and one click turns your week into a shareable recap card.
5. Granular controls
   Toggle Malware/Phishing, Advertising, and Tracking independently. Pause blocking on a single site if it breaks. Custom block/allow lists and per-country blocking included.
6. Community protection
   One click reports a phishing site for review — approved reports reach every Zevr Guard user through the daily threat feed. Reports send the domain name only, and only when you choose to.

━━━ ABOUT THE SCREENSHOTS ━━━

Screenshots use seeded sample data so the popup, side panel, and warning screen can be shown clearly in a single image. The "Dangerous" risk score shown for zevrhq.com (our own site) is a demonstration value — in real use, every score is computed from the actual third-party connections seen on that page. zevrhq.com itself is not classified as malicious; it's used only as a placeholder host.

━━━ HOW IT WORKS ━━━

• Threat list: refreshed daily from URLhaus and curated malware/phishing feeds
• Tracker ID: 115,000+ signals merged from leading open-source tracker datasets
• Geo / ASN lookup: GeoLite2, bundled and fully offline
• Matching: 100% local, via the browser's declarativeNetRequest API
• No URL uploads. No telemetry. No account required.

━━━ PRIVACY ━━━

Zevr Guard never sends your browsing history anywhere. The only automatic outbound request is a daily fetch of threat-list updates from our CDN — it carries no information about you or the sites you visit. The single exception is one you control: clicking "Report as phishing" sends that one domain name to us for review, and nothing else. Don't take our word for it: the code is public.

━━━ OPEN SOURCE ━━━

Source code (GPL-3.0): https://github.com/krystasis/zevr-guard