AI 产品解读Automatically detect and blur NSFW/haram (impermissible under Islamic values) images and videos in real-time using on-device machine learning, and block inappropriate websites
Helps Muslims and privacy-conscious users uphold the Islamic principle of lowering the gaze by automatically blurring or hiding impermissible visual content (revealing images/videos) during everyday browsing, reducing exposure to fitnah (temptation)
1. Extension content script runs on all pages at document_start, scanning images and videos. 2. On-device ML models (TensorFlow.js) detect human figures and classify gender. 3. Based on user settings (gender selection, strictness, blur amount), detected content is automatically blurred, hidden, or replaced with a solid color. 4. Declarative net request rules block known haram/NSFW websites. 5. User can configure all settings via popup or options page, including whitelisting sites and adjusting detection sensitivity. 6. Optional: Sign in with Google or email to unlock premium features (password protection, companion mode, hardening controls).
Real-time AI-powered detection and blurring of haram/NSFW images and videos on any websiteGender-specific detection (blur male, female, or both)Automatic website blocking for inappropriate/haram sitesCustomizable blur intensity, strictness, and visual style (blur, grayscale, solid color)Whitelist specific websites from blurringReport false positives/negatives via context menu
- 目标用户
- Muslim users who want to uphold Islamic gaze protection principles / Parents seeking to protect children from inappropriate content / Privacy-conscious users who prefer to avoid NSFW content while browsing / Users seeking accountability tools for content consumption
- Exact pricing/subscription details not visible in bundled code — the extension uses its own backend API (haramblur-api.md-alganzory.workers.dev) but billing flow details are not in the source package
- Cannot verify if 'free trial' converts to a paid subscription or remains free indefinitely
- The on-device ML processing is confirmed by the presence of TensorFlow.js WASM binaries and model files, but exact model capabilities cannot be assessed from binary files
