TruthLens

Fake News Detection Assistant Ecosystem

TruthLens is a decentralized, AI-powered system designed to combat misinformation through advanced forensic analysis, swarm intelligence consensus, and a decentralized trust infrastructure. By blending multi-agent LLM systems with Web3 reputation and governance mechanisms (Shardeum), TruthLens provides irrefutable fact-checking directly to end users.

🏗️ Architecture Design & Ecosystem

TruthLens operates through a multi-faceted ecosystem:

graph TD
    subgraph "Client Interfaces"
        Ext[Chrome Extension/Side Panel]
        Bot[Discord / WhatsApp Bot]
        Web[React PWA & Dashboard]
    end

    subgraph "Core Backend Services"
        API[FastAPI Gateway]
        Orchestrator[Swarm Orchestrator Agent]
        ImageAgent[Image Forensic Agent]
        SearchAgent[Web & Content Search Agent]
        Consensus[Consensus Agent]
    end

    subgraph "Infrastructure & Data"
        DB[(PostgreSQL + pgvector)]
        Redis[(Redis Cache)]
    end

    subgraph "Blockchain & Governance (Shardeum Mezame)"
        TruthDAO[TruthDAO Smart Contract]
        ReputationSBT[Reputation SBT]
    end

    Ext -->|WebSocket/REST| API
    Bot -->|REST| API
    Web -->|User Mgmt & Quests| API
    
    API --> Orchestrator
    Orchestrator --> ImageAgent
    Orchestrator --> SearchAgent
    Orchestrator --> Consensus
    
    Orchestrator <--> DB
    Orchestrator <--> Redis
    
    Consensus -->|On-chain verification| TruthDAO
    TruthDAO -->|Issue/Revoke| ReputationSBT

Loading

🌟 Key Features

• Multi-Agent Swarm Intelligence: Uses specialized AI agents for image forensics, deep-web research, and consensus to systematically evaluate truth claims.
• Decentralized Governance: Integrated natively with the Shardeum Mezame testnet. Users can vote on contested facts and build on-chain reputation via Soulbound Tokens (SBTs).
• Browser Extension (Clean Hub): A sleek Chrome side-panel providing in-context, on-page truth verification, MetaMask integration, and seamless Web3 action flows.
• Omnichannel Sentinels: Discord and WhatsApp bots proactively monitor, verify, and broadcast real-time fact-checks across heavily targeted social communities.
• Knowledge Graph Visualizations: Visual node-edge mappings of sources, citations, and related entities for transparent provenance tracking.

🚀 Running Steps

Prerequisites

• Node.js (v18+)
• Python (v3.10+)
• Docker & Docker Compose
• MetaMask (configured for Shardeum Testnet)

1. Start Infrastructure Dependencies

TruthLens backend utilizes PostgreSQL (with pgvector for semantic search) and Redis for quick caches.

# In the project root
docker-compose up -d

2. Configure Environment variables

Set up your .env file in the backend/ and root directories based on provided templates. Required API Keys: Google Gemini (GenAI), SerpAPI, Firecrawl, and your Web3 Wallet Private Key for deployments.

3. Run Backend Services

cd backend
python -m venv venv
# Windows
venv\Scripts\activate
# Mac/Linux: source venv/bin/activate

pip install -r requirements.txt
python -m uvicorn api.main:app --host 0.0.0.0 --port 8000 --reload

4. Smart Contracts Build & Deploy (Optional)

Deploy the core DAOs and token infrastructure to Shardeum Mezame testnet:

npm install # from project root
cd contracts
npx hardhat compile
python scripts/deploy_truthlens.py

5. Start Frontend & Extension Environments

Run the entire frontend ecosystem concurrently:

# To run the web dashboard
npm run dev:web

# To run the Chrome Extension
npm run dev:ext

To install the extension, navigate to chrome://extensions/, enable "Developer mode", and load the extension/dist build folder.

🤝 Community & Governance

TruthLens is governed by its users. Dispute resolutions and knowledge additions are processed on-chain using TruthDAO. All community members are incentivized to maintain high standard reputation scores to influence overall multi-agent consensus weights.