How Trunod Works
A decentralized reputation system powered by Bayesian statistics, community validation, and real-time analysis.
1. Trust Analysis
Uses Bayesian probability to calculate trust scores. Evidence is weighted: vote ratio, account age, network signals, and Sybil detection.
2. Community Validation
Claims are validated by expert juries selected from the community. Transparent voting and consensus mechanisms ensure accuracy.
📊 The Trust Score System
Trunod uses probability theory to calculate realistic trust scores. This mathematical approach is more accurate than simple weighted averages because it accounts for uncertainty and updates beliefs as new evidence emerges.
How It Works:
- 1.Prior Belief: Start at 50% trust (neutral for new users)
- 2.Collect Evidence: Analyze votes, account age, network connections
- 3.Weight Evidence: Strong signals (vote ratio) carry more weight than weak ones (age)
- 4.Update Belief: Bayesian formula combines prior with new evidence to produce final score
- 5.Detect Threats: Sybil detection identifies fake accounts and bot networks
🎬 GIF Animation: Trust Score Calculation
Visual showing: Data → Evidence → Bayesian Formula → Score
🔍 Evidence Signals
Each signal contributes to the final trust score. The more positive signals, the higher the trust rating.
Vote Ratio
Upvotes vs downvotes. High upvote ratio suggests trustworthy contributions.
Account Age
Older accounts tend to be more trustworthy. New accounts start neutral.
Network Signal
Connections to other trusted users increase credibility. "Trust by association."
Sybil Detection
AI-powered detection of fake accounts, bots, and coordinated networks.
🎬 GIF Animation: Evidence Signals in Action
Visual showing: Each signal being evaluated and weighted
⚖️ Community Jury System
When claims are disputed, Trunod assembles juries of community experts to validate or invalidate them. This decentralized approach prevents central authorities from manipulating truth.
Jury Selection Process:
- 1.Claim Disputed: A user or system flags a claim as potentially inaccurate
- 2.Topic Match: Jurors with expertise in the claim's topic (Tech, Finance, Science) are identified
- 3.Random Selection: Jury members randomly selected to prevent bias
- 4.Blind Voting: Jurors vote without knowing other votes (prevents groupthink)
- 5.Consensus & Reward: Majority verdict wins. Jurors who voted with consensus earn rewards
🎬 GIF Animation: Jury Selection & Voting Process
Visual showing: Claim → Expert Selection → Voting → Consensus
💡 Key Benefits:
- ✅ No single authority controls truth
- ✅ Expert opinions weighted more heavily
- ✅ Blind voting prevents mob mentality
- ✅ Transparent, auditable decision history
- ✅ Rewards honest validators, penalizes bad actors
Trust Score Ranges
Very Low
High risk, suspected scammer
Low
Caution advised
Neutral
Insufficient data
Good
Generally trustworthy
Excellent
Highly trustworthy
Ready to see trust in action?
Install the Trunod extension and start viewing real-time trust scores on your favorite platforms.
Get Started