Agent Ready

Agent Ready developer documentation

REST API, MCP server, authentication, OpenAPI spec, installable skills, and the methodology behind every check. Start with the quickstart, or jump to any surface below.

Last updated

Start here

REST API

MCP server

Agent Skills

Discovery and well-known files

Reference

Frequently asked questions

Where do I start as a developer?
Open the Quickstart — it walks you from API-key issuance to your first scan in under 60 seconds, with worked examples in curl, Node, Python, and MCP. From there, the Authentication guide covers how agents discover and use credentials, and the API & integrations page covers the full REST surface.
Should I use the REST API or the MCP server?
Both are first-class. Use the REST API directly from CI, scripts, or any HTTP client — install the `agent-ready-api` skill for a self-contained Agent Skill that wraps the workflow. Use the MCP server inside MCP-capable clients (Claude Desktop, Cursor, Cline, Continue, Goose) for tool-native invocation — install the `agent-ready-mcp` skill or run `npx -y agent-ready-mcp@latest`. The same Bearer key works for both surfaces.
Where is the OpenAPI spec?
OpenAPI 3.1 at https://agent-ready.dev/api/v1/openapi.json (aliased at /openapi.json). Use it with openapi-typescript, openapi-generator, or swagger-codegen to produce a typed client.
What does Agent Ready check?
Around 60 checks across the Vercel Agent Readability Spec, the llmstxt.org standard, and the agent-protocol manifests — MCP server cards (SEP-1649), A2A agent cards, agents.json, agent-permissions.json, UCP, x402, and NLWeb. See the Methodology page for every check, weight, and rating band.
Is API access free?
The /api/v1/ask endpoint (NLWeb search over Agent Ready's docs) is public — no key required. Every other endpoint requires an Agent Ready Pro account. Free-tier users can run scans from the homepage (3 per IP per 30 days anonymous, 10 per 30 days signed in) but can't mint API keys.