> For the complete documentation index, see [llms.txt](https://deepnode.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://deepnode.gitbook.io/docs/platform/trust-and-reputation-system.md).

# Trust & Reputation System

How participants build and lose trust, and how this influences task allocation, governance, and rewards.

Trust is at the core of DeepNode’s decentralized AI network. Since the system is open to anyone, DeepNode relies on a transparent, on-chain Reputation System to measure performance, reward high-quality contributors, and protect against abuse.

Each model, node, and validator accumulates a dynamic reputation score based on real activity, not promises or self-claims.

***

#### **Why Reputation Matters**

In a decentralized AI marketplace, reputation enables:

* Safer model usage — Users can choose reliable, validated models
* Better routing — The system prefers trusted nodes and validators
* Fairer incentives — Participants with higher reputation earn more
* Permissionless access — Anyone can join, but quality is earned

***

#### **Tracked Reputation Objects**

Reputation is calculated for the following on-chain entities:

* **Models**\
  Measured by success rate, request volume, feedback scores, and validator history.
* **Nodes**\
  Scored based on uptime, error rates, request handling speed, and validation outcomes.
* **Validators**\
  Evaluated through alignment with consensus, report quality, and timely execution.

Each entity is represented by an NFT (ERC-1155), and its reputation is tied directly to that NFT.

***

#### **How Reputation is Built**

Each successful action contributes to reputation:

| **Action**                               | **Reputation Impact**     |
| ---------------------------------------- | ------------------------- |
| Model returns valid output               | + Model reputation        |
| Node processes request without error     | + Node reputation         |
| Validator aligns with consensus          | + Validator reputation    |
| User provides positive feedback          | + Model & Node reputation |
| Validator submits invalid report         | – Validator reputation    |
| Node fails task or crashes               | – Node reputation         |
| Model has repeated errors or abuse flags | – Model reputation        |

Scores are updated continuously and stored on-chain or in verifiable off-chain proofs.

***

#### **Reputation Weighting**

Reputation is not just a score, it influences protocol-level decisions:

* Task routing — High-reputation nodes are prioritized for deployments
* Reward multipliers — Participants with stronger reputation earn more $DN
* Governance weight — Higher reputation can increase proposal influence
* Exit penalties — Low-reputation actors may face slashing or removal

***

#### **Sybil Resistance & Identity**

Because participation is permissionless, DeepNode:

* Requires stake-backed roles (Nodes, Validators, Creators)
* Anchors actions to NFTs
* Ties reputation to long-term identity rather than disposable wallets

This makes spamming costly and rewards building a strong identity over time.

***

#### **Reputation Transparency**

* All scores and history are publicly visible via the Dashboard
* Users can inspect reputation before staking, delegating, or interacting
* Reputation is auditable and non-resettable (except on NFT burn)