Trunod: Decentralized Credibility & Reputation Layer

*Some info below might not be up to date.

Executive Summary

The internet has fundamentally transformed how humanity communicates, learns, and forms beliefs. However, modern digital platforms optimize primarily for engagement, virality, and retention — not credibility. As a result, misinformation, manipulation, synthetic media, and performative expertise now spread at an unprecedented scale.

Trunod is a browser extension designed to bring objective credibility, expert consensus, and reputation tracking to the modern web. By acting as a universal overlay on top of major social networks (Reddit, X, YouTube, LinkedIn etc) and news websites, Trunod calculates real-time Trust Scores for digital identities and content. Trunod focuses on a user’s digital identity and online reputation to help people determine who to trust and whether information is credible. Rather than analyzing content alone, Trunod evaluates the individuals behind the information — including their behavior, expertise, and reputation — to help you make your own decisions about how to use that information. The goal is to combat misinformation and incentivize truth through a transparent, decentralized reputation system.

1. Core Architecture & How It Works

The Trunod ecosystem consists of a lightweight browser extension and a Web UI. The Trunod browser extension serves as the primary interface between users and the credibility network.

The extension injects lightweight, non-intrusive overlays into supported websites, allowing users to view contextual trust signals directly within their browsing experience.

Supported environments include:

  • social media platforms
  • discussion forums **
  • news websites
  • blogs **
  • knowledge-sharing platforms **

Rather than disrupting browsing behavior, Trunod augments existing interfaces with additional contextual information.

1. In-Browser social feeds discovery:

The extension injects non-intrusive UI elements (Discovery Modals) next to usernames on supported social media platform feeds, allowing users to instantly view the credibility of the person they are interacting with.

2. The Article Layer:

Trunod extends beyond user profiles to analyze the credibility of entire articles dynamically. When a user visits a supported news or blog site, Trunod's Article Widget pops up to provide:

  • Bias Detection: Analyzes the sentiment and political/social leaning of the text, ideological asymmetry. The objective is not ideological suppression, but transparency regarding framing patterns.
  • Opinion recognition: The system distinguishes factual reporting, analysis, opinion content, and speculative claims. This helps users contextualize whether a piece is informational or interpretive.
  • Misinformation: Compares the article's headline to its actual content to detect sensationalism and misleading titles.
  • Source Trustworthiness: Uses natural language processing (NLP) to assign a credibility score to the publisher based on historical accuracy, factual density, and content structure. Publisher credibility is evaluated using: historical factual consistency, correction frequency, editorial transparency, citation behavior, and corroboration reliability. These metrics contribute to a dynamic publisher credibility profile.
  • Citation credibility: The platform evaluates citation density, source diversity, reference quality, and corroboration patterns. Articles with weak sourcing or unsupported claims receive lower confidence scores.

3. The Reputation Score (Bayesian Engine)

At the heart of Trunod is a sophisticated Bayesian scoring engine that converts qualitative behavior into a quantifiable trust score.

  • The Bayesian Prior: Every new, unverified user* starts at a mathematically neutral 50% baseline. New to Trunod at the time of Trunod’s launch, and if the user has no data that can be used for calculation. (Meaning that if you are an early user of the Trunod, you might see that every username on most platforms has a 50-trust score).
  • Evidence Weighting: As users make claims, their score adjusts. Accurate claims serve as positive evidence (boosting the score); inaccurate or disproven claims serve as negative evidence. Positive evidence includes historically accurate claims, high-quality contributions, constructive engagement, verified expertise, and successful dispute outcomes. Negative evidence includes misinformation propagation, coordinated manipulation, toxic behavior, repeated factual inaccuracies, and malicious reporting behavior.
  • Contextual Adjustments: Trunod does not treat credibility as a global absolute. Scores are adjusted dynamically based on the domain (e.g., Finance, Science). Toxic behavior incurs heavy penalties (up to -30 points), while constructive behavior is rewarded. Users accumulate domain-specific credibility vectors across topics. This prevents false authority transfer between unrelated domains. For example, a highly credible software engineer may not automatically receive high credibility in medicine or economics.
  • Freshness Decay: There is an opportunity for accounts to undergo a gradual decay after a 12-month grace period, ensuring that high scores reflect current, ongoing behavior.

4. Voting & Jury Consensus

Trunod relies on decentralized expert consensus rather than centralized moderation to determine the truthfulness of a claim.

  • Jury Selection: When a claim is made, Trunod's graph database selects a jury of experts based on their historical accuracy, expertise verification in that specific topic taxonomy (e.g., Medicine, Law, Cybersecurity, biology, etc.).
  • Influence Weighting: Not all votes are equal. A juror's voting power is weighted by a combination of their overall Trust Score, verified expertise, institutional backing, and peer vouches.
  • Consensus Resolution: Once the jury reaches a consensus, the claim is marked as Accurate or Inaccurate, and reputation points are automatically distributed to the author and the correct jurors.

5. Disputes & Fact-Checking

Users can challenge claims they believe are false, ensuring self-correction within the ecosystem.

  • Contextual Disputes: Filing a dispute requires a logical reasoning statement and a supporting evidence URL.
  • Severity & Penalties: To prevent weaponized reporting, Trunod penalizes bad actors. If a user files a frivolous dispute and the jury rules against them, their own Trust Score suffers a severe "Failed Dispute Penalty."

6. Becoming an Expert

Expertise on Trunod is earned and verified through multiple rigorous channels:

  • Institutional Verification: Users authenticated via known institutional email domains receive high-confidence expert status.
  • Platform Verification: Users who authenticate via LinkedIn via OAuth receive verified professional status.
  • Peer Vouching: Users can vouch for one another's expertise in specific topics. Accumulating a critical mass of vouches elevates a user to Expert status.
  • Performance: Consistently voting correctly on juries or posting accurate claims organically builds a user's "Expertise Score" in their active topics.

Conclusion

Trunod creates a transparent, resilient, and community-driven ecosystem where truth is financially and socially incentivized, and expertise is verifiably earned. By combining graph databases, Bayesian mathematics, and seamless browser integration, it serves as the ultimate credibility layer for the internet.