AI Website Systems Standards & Architecture

Technical specifications and standards for building AI-native, machine-readable websites.

Current Version: Digital Karma Web Federation v6.1 | Updated: 2026-01 | Status: Stable

📋 Quick Reference

Required Endpoints

  • /ai/manifest.json
  • /ai/health.json
  • /ai/catalog.json
  • /ai/karma.json

Optional Endpoints

  • /ai/federation.json
  • /ai/datasets/*.json
  • /ai/llm.json
  • /llm.txt

Content-Type

  • All JSON: application/json
  • Structured data: application/ld+json
  • CORS: enabled
  • Encoding: UTF-8

Timestamps

  • Format: ISO 8601
  • Timezone: UTC
  • Example: 2026-01-15T10:30:00Z

🌐 Federation Protocol v6.1

The Digital Karma Web Federation enables AI-ready websites to discover, understand, and trust each other through standardized endpoints and transparent reputation scoring.

Core Principles

  1. Discoverability — Sites expose predictable endpoints at /ai/*
  2. Interoperability — All endpoints use standard JSON formats
  3. Transparency — Quality scores are calculated openly
  4. Decentralization — No central authority required
  5. Versioning — Protocol versioning for backward compatibility

Compliance Levels

Level Requirements Badge
Basic manifest.json + health.json Federation Member
Standard + catalog.json Federation Compliant
Full + karma.json + federation.json Federation Complete
Elite + datasets + automation Federation Elite

Read Full Federation Spec →

📄 Manifest Specification (/ai/manifest.json)

The manifest.json file is your site's identity card — the first file AI agents query to understand what your site is and what it offers.

Required Fields

{
  "name": "string",              // Site name
  "description": "string",       // Short description
  "url": "string",              // Canonical URL
  "federation_version": "string", // e.g., "6.1"
  "last_updated": "ISO8601",    // UTC timestamp
  "contact": {
    "email": "string",
    "website": "string"
  },
  "ai_endpoints": [              // Array of AI endpoint URLs
    "/ai/health.json",
    "/ai/catalog.json",
    "/ai/karma.json"
  ]
}

Optional Fields

Read Full Manifest Spec →

💚 Health Endpoint (/ai/health.json)

Real-time system health and operational metrics.

Required Fields

{
  "status": "healthy|degraded|down",
  "last_check": "ISO8601",
  "uptime_percent": 0.0-100.0,
  "response_time_ms": number,
  "metrics": {
    "total_pages": number,
    "total_datasets": number,
    "last_content_update": "ISO8601"
  }
}

Health Status Definitions

Read Full Health Spec →

📚 Catalog Endpoint (/ai/catalog.json)

Machine-readable inventory of all content and resources on the site.

Structure

{
  "site_url": "string",
  "generated_at": "ISO8601",
  "total_items": number,
  "categories": [
    {
      "name": "string",
      "items": [
        {
          "id": "string",
          "title": "string",
          "url": "string",
          "type": "string",        // e.g., "website", "tool", "service"
          "updated": "ISO8601",
          "schema_type": "string"   // Schema.org type
        }
      ]
    }
  ]
}

Read Full Catalog Spec →

⭐ Digital Karma Scoring v2.1

Transparent, automated quality and trust scoring for AI websites.

Scoring Formula

Total Score = (Schema Coverage × 0.20) + (Content Freshness × 0.15) + (AI Endpoints × 0.25) + (Federation Presence × 0.15) + (External Links × 0.10) + (Technical Quality × 0.10) + (Dataset Quality × 0.05)

Signal Breakdown

Signal Weight Measurement
Schema Coverage 20% Presence and quality of Schema.org markup
Content Freshness 15% Last update timestamp, update frequency
AI Endpoints 25% Required endpoint presence and completeness
Federation Presence 15% Network participation, bidirectional links
External Links 10% Quality and relevance of outbound links
Technical Quality 10% Performance, security, accessibility scores
Dataset Quality 5% Structured dataset availability and validity

Badge Thresholds

Read Full Scoring Spec →

🔗 Schema.org Implementation Standards

Structured data markup requirements for AI discovery and understanding.

Required Entity Types

Recommended Entity Types

Implementation Format

Use JSON-LD in <script type="application/ld+json"> tags. Place in <head> or end of <body>.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "AI Website Systems",
  "url": "https://www.aiwebsitesystems.com",
  "description": "Directory and knowledge base for AI-native website systems"
}
</script>

Read Full Schema Standards →

📝 LLM.txt Standard

Plain text file format optimized for LLM consumption and site understanding.

Location

/llm.txt at site root (alongside robots.txt)

Structure

# Site Name
> One-line description

## About
Brief overview of the site's purpose and content

## Core Content
- Homepage: /
- Primary section: /section/
- Key resource: /resource/

## AI Endpoints
- Manifest: /ai/manifest.json
- Health: /ai/health.json
- Catalog: /ai/catalog.json

## Datasets
- Main dataset: /datasets/data.json

## Context for LLMs
Additional context that helps LLMs understand your site

Read Implementation Guide →

💾 Dataset Standards

Structured data file formats for AI consumption.

Recommended Format

JSON with Schema.org Dataset metadata

Required Metadata

Storage Location

/datasets/ or /ai/datasets/

Read Full Dataset Spec →

✅ Implementation Checklist

Use this checklist to ensure full federation compliance.

Level 1: Basic Compliance

Level 2: Standard Compliance

Level 3: Full Compliance

Level 4: Elite Compliance

🔧 Validation Tools

Tools for validating your AI website system implementation.

View All Developer Tools →

Have Feedback on These Standards?

The Digital Karma Web Federation standards are community-driven. Suggest improvements or report issues.

Contact Us