What is agent.json?

agent.json is a JSON file placed at /.well-known/agent.json that describes your website or service to autonomous AI agents. It includes your service name, description, capabilities, API endpoints, and contact information — everything an AI agent needs to decide whether and how to interact with your site.

Think of it as a machine-readable “about us” page. While llms.txt provides content context and robots.txt controls access, agent.json describes capabilities — what actions an AI agent can take on your site.

Example agent.json

{
  "name": "Score for AI",
  "description": "AI readiness scanner - analyzes websites across 48 checks",
  "url": "https://scoreforai.com",
  "capabilities": {
    "scan": {
      "description": "Scan a website for AI readiness",
      "endpoint": "/api/scan",
      "method": "POST",
      "parameters": {
        "url": { "type": "string", "required": true }
      }
    }
  },
  "contact": {
    "email": "[email protected]"
  },
  "documentation": "/api/docs"
}

Why Implement agent.json?

As AI agents evolve from simple chatbots to autonomous agents that can take actions on behalf of users, they need a standard way to discover what services are available. agent.json fills this role by:

• Enabling autonomous discovery — AI agents can find your capabilities without human intervention
• Describing API endpoints — Agents learn how to interact with your service programmatically
• Building trust signals — Having agent.json shows you're prepared for the AI-first web
• Improving AI readiness score — It's a direct check in the Score for AI scanner

agent.json vs Other Standards

How to Implement

Implementation takes about 10 minutes:

1. Create a .well-known directory in your site's public/root folder
2. Add an agent.json file with your service description
3. Include at minimum: name, description, url, and contact
4. Optionally add capabilities, API endpoints, and documentation links
5. Scan your site to verify it's detected