🏷️
5 Claim Type Classifications
Categorizes posts into Official Fact, Personal Opinion, Breaking News, Rumor/Leak, or Satire/Hoax with distinct color badges inserted above the tweet body.
⚡
Viewport-Only Lazy Evaluation
Powered by IntersectionObserver, only evaluating tweets that scroll into view. If you fast-scroll past tweets, evaluation is automatically cancelled to eliminate wasted API calls.
🛡️
Label-Based Text Folding
Toggle off unverified rumor or satire categories in the popup. Tweets matching those categories fold away their text and images automatically until clicked.
📊
Local Timeline Statistics
Track total verified posts, category breakdown percentages, cache hit rate, and average latency directly in your browser. Review "Today" and "All Time" metrics anytime.
🔒
Direct Transmission & Zero Developer Tracking
Verification requests go directly from your device to TypeSafe AI without passing through or being stored on developer servers. Settings and statistics stay strictly in your local sandbox.
🌐
Full Bilingual Support
Badges, popovers, settings, and documentation seamlessly support both Japanese and English, categorizing global tweets accurately.