De Slop

De-Slop - v3.1.0
Detect and remove AI-generated slop from web pages. 600+ patterns, 11 languages, 8 analysis tools, 3 themes. Zero API calls.

Detailed Description
De-Slop automatically detects and removes AI-generated filler content from web pages. 600+ detection patterns across 11 languages, 8 built-in analysis tools, and three visual themes – all running entirely in your browser with zero external API calls.

Core Detection
Three-tier weighted pattern matching (600+ patterns)
11 languages: English, Spanish, French, German, Italian, Portuguese, Swedish, Polish, Japanese, Korean, Chinese
Real-time scanning of dynamically loaded content (infinite scroll, AJAX)
Badge counter showing blocked slop on the current page
Configurable sensitivity slider (1-5)
Detection Only Mode – highlight slop instead of removing it, with hover tooltips
Custom pattern editor – add your own patterns with custom weights
Export/import pattern sets
Analysis Tools
8 built-in tools for writing analysis and slop detection:

Slop Checker – paste text and see what gets flagged with real-time highlighting and match explanations
Humanize Score – 8-metric writing quality analysis with authenticity scoring
Batch Checker – score multiple text blocks at once with CSV export
Text Rewriter – inline fix suggestions for detected slop phrases
Text Comparator – side-by-side draft comparison with score deltas
URL Analyzer – fetch and score any webpage for slop content
Slop Machine – gamified slot machine for learning better word choices (90+ slop/better pairs)
Pattern Tester – validate detection patterns across all 11 languages
Themes
Three visual themes synced across all pages:

Dark (default) – comfortable dark interface
Light – clean light interface for bright environments
Midnight – true black OLED theme for maximum contrast
Cycle themes from the popup header or select from Settings. Your choice persists across pages and browser restarts.

LinkedIn Fixer
When on LinkedIn.com, additional features activate:

Block auto-play videos
Block ad tracking
Darker mode – true black LinkedIn theme

Advanced Features
Emoji slop detection – 166 patterns for emoji spam and engagement bait
Stop-word filtering – detect “I’m excited to announce” style posts

Em dash detector – AI loves em dashes
Political content filter – optional political content detection

YouTube filtering – video titles, descriptions, Shorts, and clickbait detection
Improvement suggestions – better alternatives for sloppy phrases

Detection System
De-Slop uses a three-tier aggression system:

Tier 1: AI-Specific Indicators (3 points each)
Signature AI phrases and patterns:

“delve into”, “navigate the landscape”, “tapestry of”, “realm of possibilities”
Em dash overuse, excessive transitions
“In today’s fast-paced world…”, “As we continue to evolve…”
Stop words: “I’m excited to announce”, “Thrilled to share”
Listicle patterns: “Top 10 ways to…”, “Ultimate guide to…”
Active at all sensitivity levels.

Tier 2: Corporate Buzzwords (2 points each)
LinkedIn’s greatest hits:

“synergy”, “leverage”, “circle back”, “low-hanging fruit”
“take this offline”, “touch base”, “thought leadership”
“AI-powered”, “blockchain-enabled”, “digital transformation”
Active at sensitivity 3+.

Tier 3: Marketing Spam (1 point each)
The widest net:
“free”, “guaranteed”, “limited time offer”, “buy now”
“amazing”, “incredible”, “revolutionary”, “mind-blowing”
“basically”, “essentially”, “actually”
Active at sensitivity 4+.

Scoring Thresholds
1 Very conservative – obvious AI only
2 Conservative
3 AI + corporate
4 Aggressive – includes marketing
5 Nuclear – maximum detection

How It Works
Scans page content using platform-specific selectors
Assigns weighted “slop score” based on pattern matches
Removes elements exceeding threshold (or highlights in Detection Only Mode)
Updates badge counter
Monitors for dynamic content (infinite scroll, AJAX)
Platform Support
LinkedIn: individual posts
Twitter/X: tweets
Medium: articles
YouTube: videos, Shorts, comments, clickbait detection
All sites: generic article/post/card selectors

Privacy
Runs entirely locally – no data sent to servers
Stores settings in Chrome sync storage
No analytics or tracking
No external API calls

Open source under GPL-3.0
Permissions
storage – save your settings
activeTab – access current page content
scripting – dynamic language pattern injection
sidePanel – sidebar interface
declarativeNetRequest – block LinkedIn ads (optional)

Installation
Install from the Chrome Web Store or load unpacked for development: https://github.com/HxHippy/DeSlop

Version History
v3.1.0 – Theme System & About Page
Three-theme system: Dark (default), Light, and Midnight (OLED true black)
CSS variable architecture across all 11 UI pages
Theme cycle button in popup, theme selector in settings
Theme persists across pages and browser restarts
About page with project info and privacy statement
Improved button contrast for accessibility

v3.0.0 – Multi-language & Tools Suite
11-language pattern system (en, es, fr, de, it, pt, sv, pl, ja, ko, zh)
Full i18n with localized UI
6 new analysis tools (humanize, batch, rewriter, compare, url-analyzer, pattern tester)
Externalized pattern architecture with shared scoring engine
Expanded settings with pattern management overhaul

v2.7.2 – Political Content Filter & YouTube
Political content filter
YouTube video/Shorts/comment filtering
Clickbait pattern detection
Improvement suggestions

v2.4.4 – Detection Only Mode
Highlight-only mode (non-destructive detection)
Emoji detection expanded to 166 patterns
Simplified UI with sensitivity-driven tier control

v2.0.0 – LinkedIn Fixer & Dark Theme
Complete UI redesign with dark theme
Pattern customization system with import/export
LinkedIn video blocking, ad blocking, darker mode

v1.0.0 – Initial Release
Three-tier detection system
Basic popup controls

Credits
Made by Kief Studio (https://kief.studio) | Developed by HxHippy & Meelie
- hxhippy.com
- meelie.art

License
GNU General Public License v3.0 (GPL-3.0)