Transform your X feed. Filter AI slop, engagement bait, and bot content. See what's real.

Feed Cleaner scores every post on your X timeline across 4 dimensions:

Posts get graded A through F with full transparency — click any score to see exactly why.

• Three Filter Modes: Highlight (color-code), Dim (fade low quality), Clean (remove entirely)
• Quality Threshold Slider: Set your own standards
• Gem Detection: High-quality posts get a 💎 badge so you don't miss them
• Detail Breakdown: Click any score for a full analysis with color-coded category bars
• Session Stats: Posts scanned, filtered, time saved — right in the popup
• Account Audit Dashboard: See your best and worst follows ranked by content quality
• Share Your Score: Copy your feed health score and share it

100% local. Always.

• All analysis runs in your browser — zero API calls
• We never see your feed, your scores, or your data
• No tracking, no analytics, no telemetry
• No account required

(Coming soon)

1. Clone this repo
2. npm install && npm run build
3. Open chrome://extensions/
4. Enable "Developer mode"
5. Click "Load unpacked" → select the dist/ folder

npm install
npm run dev          # Vite dev server
npm run build        # Production build
npx vitest run       # Run tests (28 passing)
npx tsc --noEmit     # Type check

The detection engine is adapted from Humanizer, extended for social media:

• Text Statistics: Flesch-Kincaid readability, burstiness (sentence length variance), type-token ratio
• Pattern Matching: 100+ regex patterns across AI vocabulary, bait hooks, bot behaviors
• Weighted Scoring: Each detector contributes to a weighted overall score (configurable)
• Confidence Scaling: Short texts get lower confidence to avoid false positives

src/
├── detection/          # Core scoring engine
│   ├── ai-detector     # AI content detection (Humanizer port)
│   ├── bait-detector   # Engagement bait patterns
│   ├── bot-detector    # Bot/farm account signals
│   ├── originality     # Quality/originality scoring
│   ├── scorer          # Master scorer combining all detectors
│   ├── text-stats      # Statistical text analysis
│   └── patterns        # All detection patterns and vocabulary
├── content/            # Content script (injected into X)
│   ├── scanner         # DOM observation and tweet extraction
│   ├── overlay         # Visual indicators (halos, badges, panels)
│   └── filter          # Highlight/dim/clean filter logic
├── popup/              # Extension popup UI
├── dashboard/          # Full analytics dashboard
├── background/         # Service worker (stats, settings)
└── shared/             # Types, constants, storage helpers

• 2,600+ lines of TypeScript
• 28 tests passing
• 100+ detection patterns
• 500+ AI vocabulary terms
• Zero runtime dependencies (React only for popup/dashboard UI)

MIT

Built by @brandonwise