@aiendpoint/cli

Make your service AI-readable in minutes. Generate a /ai endpoint spec from OpenAPI, or build one interactively.

AI reads a webpage  ->  10,000-50,000 tokens  (mostly noise)
AI reads /ai        ->          ~800 tokens  (0% noise)

Quick Start

# From an OpenAPI spec (local file or URL)
npx @aiendpoint/cli init --openapi https://petstore3.swagger.io/api/v3/openapi.json

# Interactive wizard
npx @aiendpoint/cli init

# Validate a live /ai endpoint
npx @aiendpoint/cli validate https://api.aiendpoint.dev

Commands

init

Generate a /ai spec for your service.

npx @aiendpoint/cli init [options]

| Option | Description | |--------|-------------| | --openapi <url\|file> | Generate from OpenAPI/Swagger spec | | --framework <type> | Target framework: nextjs, fastify (auto-detected if omitted) | | --output <dir> | Output directory (default: current directory) | | --json-only | Only output ai.json, skip route handler generation | | --force | Overwrite existing files |

What it does:

1. Converts your OpenAPI spec to /ai format (or walks you through a wizard)
2. Validates and scores the result (0-100)
3. Auto-detects your framework (Next.js or Fastify)
4. Generates ai.json + a route handler file

$ npx @aiendpoint/cli init --openapi ./openapi.json

  Converted: 19 capabilities extracted

  Score: 75/100 (Good (AI-Ready))
  Capabilities: 19
  Token estimate: ~1462 tokens

  Framework: nextjs

  Files:
    + ai.json
    + app/ai/route.ts

  Done! Your service now has a /ai endpoint.

validate

Check if a live /ai endpoint is valid.

npx @aiendpoint/cli validate <url>

$ npx @aiendpoint/cli validate https://api.aiendpoint.dev

  Found: https://api.aiendpoint.dev/ai (415ms)
  Score: 85/100 (Good (AI-Ready))
  Capabilities: 6
  Token estimate: ~832 tokens

Framework Middleware

For serving /ai in production, use the framework-specific packages:

# Next.js
npm install @aiendpoint/next

# Fastify
npm install @aiendpoint/fastify

See @aiendpoint/next and @aiendpoint/fastify.

Links

• aiendpoint.dev - Registry & documentation
• Spec docs - The /ai standard
• MCP server - For AI agents
• GitHub