NameAI SDK

NameAI extends NameGuard with sophisticated natural language processing capabilities to help evaluate and understand Ethereum Name Service (ENS) names.

NameGuard is designed to inspect and prevent malicious use of ENS names by providing comprehensive security checks. See the NameGuard SDK documentation for details about the base security features.

⚠️ This SDK is BETA. Things will change based on the community feedback.

API Implementation

This SDK provides a TypeScript interface to the NameAI API. Developers interested in the NameAI API internals can find the implementation in the NameAI backend service.

Features

Additional features provided by NameAI on top of NameGuard:

• Natural Language Processing: Evaluate how well ENS names represent meaningful words or phrases
• Tokenization: Break down names into meaningful word components with probability scoring
• Purity Scoring: Assess the cleanliness and quality of ENS names
• Sort Scoring: Get relative ranking scores for ENS names

The @namehash/nameai SDK provides full type-safety when working with the NameAI API.

Install

Install NameAI via npm, yarn or pnpm:

npm install @namehash/nameai

Usage

Import nameai:

import { nameai } from "@namehash/nameai";

Basic Name Inspection

Inspect a name:

const report = await nameai.inspectName("vitalik.eth");

// Example response:
{
  nameai: {
    // Quality score (0.0 to 1.0) for the first label
    purity_score: 0.29,
    
    // Ranking score (0.0 to 1.0) for the first label
    sort_score: 0.35,
    
    analysis: {  // undefined for uninspected names
      // The normalization status of the name (normalized, unnormalized, or unknown)
      status: "normalized",
      
      // Details about the inspected name component
      inspection: {
        label: "vitalik",
        // ... other inspection details
      },
      
      // Text meaningfulness (0.0 to 1.0)
      probability: 0.95,
      
      // Natural log of probability (≤ 0.0)
      log_probability: -0.05,
      
      // Minimum words in valid tokenizations
      word_count: 1,
      
      // Recommended word breakdown (may be undefined)
      top_tokenization: ["vitalik"],
      
      // All possible tokenizations (up to 1000)
      tokenizations: [
        {
          tokens: ["vitalik"],
          probability: 0.95,
          log_probability: -0.05
        }
      ]
    }
  },
  nameguard: NameGuardReport  // Standard NameGuard security inspection
}

Tokenization Inspection

The SDK provides detailed tokenization inspection:

const { nameai: { analysis } } = await nameai.inspectName("cryptowallet.eth");

console.log(analysis.top_tokenization); // ["crypto", "wallet"]
console.log(analysis.tokenizations); // Array of possible tokenizations with probabilities

Custom Client

You can create a custom client with different settings:

import { createClient } from "@namehash/nameai";

const nameai = createClient({
  nameaiEndpoint: "...",
  network: "sepolia"
});

Limitations

• Maximum name length: 200 characters (including dots)
• Maximum unknown labels: 5
• Names exceeding these limits will have undefined analysis

Contact Us

Visit our website to get in contact.

License

Licensed under the MIT License, Copyright © 2023-present NameHash Labs.

See LICENSE for more information.