Stop guessing. Start verifying.
VeritasOS gives AI agents and dApps the power to decide who and what to trust before taking action.

AI agents are making high-stakes decisions daily, but they're flying blind:

• 🤖 Operating on untrusted data from anonymous sources
• 🎭 No defense against bots, Sybils, and disinformation clusters
• 🔗 Missing reputation backbone for critical trust decisions

Result: Your AI agent can't answer: "Should I trust this identity? Amplify this content? Transact with this address?"

VeritasOS is the decentralized reputation oracle that transforms social graphs into verifiable trust scores.

We ingest the Umanitek Guardian network, compute Sybil-resistant reputation scores, and publish them as machine-readable Knowledge Assets on the OriginTrail DKG.

• PageRank-inspired algorithm over Guardian social graph
• Normalized scores (0-1) with intuitive trust levels: High/Moderate/Low
• Bot detection through network analysis

• JSON-LD Knowledge Assets on OriginTrail DKG
• Structured scores: global, networkTrust, misinfoResistance
• Provenance tracking: algorithm version, timestamp, computation source

// Your AI agent can now query trust levels
const reputation = await trust_getReputation("@factchecker_alice");
// Returns: { score: 0.95, recommendation: "TRUST", explanation: "...", dkg_ual: "..." }

Available Tools:

• trust_getReputation(subjectId) - Get identity trust scores
• trust_getContentTrust(contentUrl) - Content credibility assessment
• Returns actionable recommendations: TRUST / VERIFY_BEFORE_TRUST / DO_NOT_TRUST

• Search identities by handle or ID
• Trust cards with visual scores and explanations
• Leaderboard ranking of top trusted identities
• DKG asset references for verification

# Query reputation by handle
GET /api/reputation/by-handle/factchecker_alice

# Get top trusted identities  
GET /api/top?limit=10

# Content trust assessment
GET /api/content-trust?url=https://...

graph TD
    A[🔗 Guardian Social Graph] --> B[🔄 Reputation Engine]
    B --> C[📊 Trust Scores]
    C --> D[🌐 OriginTrail DKG]
    
    D --> E[🔌 Access Layers]
    E --> F[🤖 MCP Tools]
    E --> G[📡 REST API]
    E --> H[🎨 Web UI]
    
    F --> I[AI Agents]
    G --> J[dApps & Protocols]
    H --> K[Humans & Analysts]
    
    

Scenario 1: Trusted Identity

Search: "@factchecker_alice"
Result: ✅ 96% Trust Score (HIGH)
Explanation: "Strong incoming endorsements from trusted network"
DKG Reference: ✅ Verifiable on OriginTrail

Scenario 2: Suspicious Identity

Search: "@bot_larry" 
Result: ❌ 12% Trust Score (LOW)
Explanation: "Isolated node with minimal trusted connections"
DKG Reference: ✅ Verifiable on OriginTrail

• Node.js 18+
• npm or yarn

# Clone and setup
git clone https://github.com/your-org/veritasos
cd veritasos

# Install dependencies
npm install

# Start backend (computes reputations & serves API)
cd backend && npm run dev

# Start frontend (in new terminal)
cd frontend && npm run dev

Access Points:

• 🌐 Web UI: http://localhost:3000
• 📡 REST API: http://localhost:3001/api
• 🤖 MCP Server: Integrated for AI agent tooling

// Before acting on user instruction, check reputation
const trust = await getReputation(userIdentity);
if (trust.recommendation === "TRUST") {
  await executeHighStakesAction();
} else {
  await requestHumanReview();
}

• Filter bot accounts from your platform
• Gate high-value actions behind reputation thresholds
• Display trust badges for credible users

• Composable trust primitive for OriginTrail/Polkadot ecosystems
• Reputation-as-a-service for trust-critical applications
• Transparent, auditable reputation backbone

"In a world of increasingly autonomous AI, VeritasOS ensures that trust becomes a verifiable, decentralized primitive rather than a centralized vulnerability."

VeritasOS isn't just another reputation system—it's the trust backbone for the next generation of AI and Web3 applications.