aiendpoint.dev
Docs

Introduction

What the /ai standard is, why it exists, and how to navigate the rest of the documentation.

Why /ai

The web already has conventions for machines:

robots.txt

Tells crawlers what NOT to do

sitemap.xml

Tells crawlers where pages are

/ai

Tells AI agents what you can DO

/ai is a lightweight JSON endpoint that tells an AI agent what a service can do and how to call it.

Instead of scraping HTML, guessing hidden APIs, or reading a long human-focused docs site, an agent can fetch one predictable endpoint and get a compact machine-readable contract.

What you get

  • A predictable GET /ai entry point
  • A compact service summary and capability list
  • Clear authentication hints
  • Token-efficiency hints such as compact mode or field filtering
  • A validation path for checking whether an implementation is usable

Documentation map

Use these pages in order if you are implementing the spec for the first time:

  1. Quick Start - generate your first /ai endpoint
  2. CLI - generate from OpenAPI, validate, scaffold
  3. Serve - one-line middleware for Express, Fastify, Next.js, Hono
  4. Spec Reference - all fields and their meaning
  5. Validation - scoring and compliance

If you are integrating AIEndpoint with clients or workflows:

  • MCP Server - let AI agents discover any website
  • Skill - Claude Code skill for inline generation

Previous

Concepts

Next

Quick Start