# Agent Ready

> Scanner that scores any website for AI agent readability against the Vercel Agent Readability Spec, llmstxt.org, and agent-protocol manifests.

Agent Ready is an independent project built and maintained by Mark Lavercombe that scores public websites on how easily AI agents can discover, parse, and act on them. It runs about 72 checks across the Vercel Agent Readability Spec, the llmstxt.org standard, and agent-protocol manifests, and returns a 0-100 score with a plain-English fix for each failing check. It is available as a web app, REST API, MCP server, CLI, SDKs, browser extensions, and editor plugins.

## What It Is

Agent Ready is an agent-readability scanner. Agent readability covers discovery files (llms.txt, robots.txt, sitemaps), structural signals (semantic headings, canonical links, structured data, markdown mirrors), and protocol manifests (MCP server cards, A2A agent cards, agents.json, agent-permissions.json). Users enter a URL and receive a per-check breakdown with remediation guidance. The score is the share of passed checks, rated from Needs Improvement to Excellent.

## What the Scanner Checks

- The Vercel Agent Readability Spec: site-wide checks plus per-page checks such as meta tags, JSON-LD, headings, markdown mirrors, and content negotiation.
- llmstxt.org: checks on the llms.txt format.
- Agent protocols: MCP, A2A, agents.json, agent-permissions.json, UCP, x402, NLWeb and others.
- Accessibility: a separate sub-score from WCAG 2.2 and layout-stability checks on the homepage, which does not affect the readability score.

## Fixing and Monitoring

The site offers generators for llms.txt, AI robots.txt, agents.json, permissions, API catalog, and A2A cards. It also offers validators for AGENTS.md, MCP cards, agents.json, UCP, ACP, and x402, plus scan comparison. Weekly monitoring is described as catching deploys that undo fixes.

## Developer Surfaces

A REST API with an OpenAPI 3.1 spec and a hosted MCP endpoint expose scan_site, get_scan, and ask tools. A CLI (agent-ready-scanner), JavaScript/TypeScript and Python client SDKs, a GitHub Action, installable Agent Skills, a Cursor plugin, a Gemini CLI extension, a ChatGPT app, and Chrome, Edge, and Firefox extensions are listed. The Cursor plugin bundles an MCP config, a skill, and a rule, and the surrounding client tooling is MIT-licensed.

## Features
- 72 checks across the Vercel Agent Readability Spec, llmstxt.org, and agent protocols
- 0-100 agent readability score with per-check breakdown and fixes
- Separate WCAG 2.2 accessibility sub-score
- llms.txt, robots.txt, agents.json, permissions, API catalog, and A2A card generators
- Validators for AGENTS.md, MCP cards, agents.json, UCP, ACP, and x402
- Scan comparison
- REST API with OpenAPI 3.1 spec
- Hosted MCP server with scan_site, get_scan, and ask tools
- CLI, JS/TS and Python SDKs, GitHub Action
- Weekly monitoring
- Public NLWeb /ask endpoint

## Integrations
Cursor, Claude Desktop, Claude Code, Cline, VS Code, Windsurf, Gemini CLI, ChatGPT, GitHub Actions, Chrome, Edge, Firefox, Model Context Protocol

## Platforms
WEB, API, BROWSER_EXTENSION, VSC_EXTENSION, DEVELOPER_SDK, CLI

## Pricing
Freemium — Free tier available with paid upgrades

## Links
- Website: https://agent-ready.dev
- Documentation: https://agent-ready.dev/docs
- Repository: https://github.com/mlava/agent-ready-cursor-plugin
- EveryDev.ai: https://www.everydev.ai/tools/agent-ready