agentrank

agentrank

The Page Speed for the Agentic Web

Measure how reliably an AI agent can navigate your website. Get an Agent Visibility Score (0-100) based on 5 signals: Permissions, Structure, Accessibility, Hydration, and Hostility.

🌐 Try it live at agentrank.it

Quick Start

# Run directly with npx
npx agentrank audit https://example.com

# Or install globally
npm install -g agentrank
agentrank audit https://example.com

CLI Usage

# Quick scan (~5 seconds)
agentrank audit https://example.com

# Verbose output
agentrank audit https://example.com --verbose

# JSON output (for scripting)
agentrank audit https://example.com --json

Programmatic Usage

import { scanUrl, getGrade } from 'agentrank';

const result = await scanUrl({
  url: 'https://example.com',
  mode: 'quick',      // 'quick' (default) or 'deep'
  timeout: 30000,     // Timeout in ms
});

console.log(`Score: ${result.agentScore}/100`);
console.log(`Grade: ${getGrade(result.agentScore)}`); // A, B, C, D, or F

// Access individual signals
result.signals.permissions.status;  // 'pass' | 'warn' | 'fail'
result.signals.structure.score;     // 0-100
result.signals.accessibility.recommendations; // string[]

Note: Deep mode (mode: 'deep') requires a self-hosted browser-use engine. See Deep Mode Setup for Docker configuration.

Agent Visibility Score

Signal	Weight	What It Measures
Permissions	20%	robots.txt analysis
Structure	25%	Semantic HTML density
Accessibility	25%	ARIA labeling & accessibility tree
Hydration	15%	Time-to-Interactive
Hostility	15%	Bot-blockers (CAPTCHA, etc.)

MCP Server (IDE Integration)

Start a local MCP server for Cursor or Claude Desktop:

agentrank mcp --port 3000

curl -X POST http://localhost:3000/mcp \
  -H "Content-Type: application/json" \
  -d '{"action": "audit", "url": "https://example.com"}'

Requirements

• Node.js 22+
• Playwright browsers (auto-installed on first run)

Links

• Website: agentrank.it
• GitHub: github.com/kiarashplusplus/AgentRank.it
• Issues: Report bugs

License

Apache-2.0 © Kiarash Adl