TruthLayer
Architect & Full-Stack Developer
Overview
Chrome Extension (Manifest V3) + Node.js Express Backend + Next.js 15 Dashboard that reveals hidden intentions of every website you visit. Powered by Claude AI (Anthropic) for intent classification and dark pattern detection. Provides Trust Score (0-100), detects 10+ dark pattern manipulation tactics (fake urgency, confirmshaming, disguised ads, forced action), audits data collection & third-party trackers, estimates AI-generated content percentage, and enables public report sharing via truthlayer.io/report/domain. Features activeTab permission (privacy-first), local 24h caching, and automated CI/CD pipeline.
The Problem
Modern websites employ sophisticated dark patterns and manipulation tactics to trick users into actions they would not willingly take — fake urgency countdowns, confirmshaming, disguised ads, roach motel patterns (easy to sign up, hard to cancel), forced actions, and hidden data tracking. Users have no way to know the true intentions of a website before engaging with it. There is no existing tool that provides a comprehensive, AI-powered analysis of website trustworthiness, manipulation tactics, and data collection practices in real-time.
The Solution
TruthLayer is a Chrome Extension + Backend API + Web Dashboard ecosystem that reveals the hidden intentions of every website. With one click, it analyzes the webpage using a multi-factor engine: rule-based client-side dark pattern detection, Claude AI-powered intent classification, and a proprietary Trust Scorer algorithm. Results include Trust Score (0-100), primary/secondary/tertiary intents, detected dark patterns with evidence, data collection audit with third-party tracker enumeration, AI-generated content estimation, and manipulation level classification. Reports can be shared publicly via truthlayer.io/report/domain.
Key Features
Trust Score Engine
Multi-factor scoring algorithm (0-100) that evaluates website trustworthiness based on dark pattern presence, data collection practices, content transparency, and AI analysis. Provides an immediate, intuitive credibility assessment.
Dark Pattern Detection
Detects 10+ manipulation tactics including fake urgency, confirmshaming, roach motel, disguised ads, forced action, social proof fabrication, and more. Each detection includes specific evidence from the page.
AI-Powered Intent Classification
Backend Claude AI analyzes webpage content to classify primary, secondary, and tertiary intents (e-commerce, news, social media, phishing, etc.) with manipulation level assessment (low/medium/high/extreme).
Data Collection & Tracking Audit
Comprehensive audit of data collection practices including third-party cookies, trackers, fingerprinting scripts, and data types harvested. Displays percentage of page weight dedicated to tracking.
AI Content Estimation
Estimates the percentage of page content generated by AI with confidence scoring. Helps users identify potentially automated or low-effort content.
Public Report Sharing
Share analysis results via public URL (truthlayer.io/report/domain.com) for community awareness and collaborative website vetting.
Architecture
Chrome Extension (Manifest V3) with popup UI, service worker background, and content script for metadata extraction. Communicates via REST API to Node.js Express backend with in-memory caching (24-hour TTL). Backend integrates Claude AI for intent classification and runs rule-based dark pattern detection engine. Web dashboard built with Next.js 15 provides landing page, about, privacy policy, and public report pages with SSR rewrites to backend.
How to Use
Install the Chrome Extension by cloning the repo and loading the extension/ folder via chrome://extensions with Developer mode enabled. Start the backend server locally with cd backend && npm install && npm run dev (requires Node.js 18+). The web dashboard runs with cd web && npm install && npm run dev. For production, use docker-compose up --build. Click the TruthLayer icon on any website to analyze it instantly. Results are cached locally for 24 hours. All analysis is private — no browsing history is tracked.
Impact & Results
Empowers users with unprecedented transparency about website intentions and manipulation tactics. The Trust Score provides an immediate, intuitive metric (0-100) for assessing website credibility. The detailed dark pattern detection educates users about manipulation techniques they previously could not identify. Claude AI integration enables nuanced intent classification beyond simple keyword matching. Public report sharing creates a community knowledge base about website practices. The extension follows Manifest V3 with privacy-first architecture — only activeTab permission, local caching, no user tracking, and HTTPS-only API communication.