What is AI agent readiness?
AI agent readiness is how prepared a website is for autonomous AI agents — the LLM-driven clients that crawl, retrieve, transact, and act on behalf of users. A ready site exposes the protocols and discovery files agents look for: MCP servers, llms.txt, payment endpoints, OpenAPI specs, and structured signals about what it offers. AgentGrade scans any URL and returns a graded report so site owners can see what agents see and close the gaps.
What AgentGrade checks
- Discovery files —
llms.txt, llms-full.txt, robots.txt, sitemap.xml, humans.txt, agent skill manifests
- MCP (Model Context Protocol) — declared and live MCP server endpoints, transport, and tool surface
- Payment protocols — x402, L402, Stripe Tempo / MPP, and other agent-native payment signals
- OpenAPI & APIs — published specs, server URLs, paid-endpoint probing, GraphQL surface
- Content negotiation — JSON/Markdown variants, agent-friendly Accept-header handling, JS-rendering risk
- Identity & standards — A2A, WebMCP, structured data, canonical metadata, security.txt
How it works
- Enter a URL and start a free, instant scan.
- AgentGrade probes the site from an agent's perspective — fetching, parsing, and verifying each declared capability.
- Get a score plus a per-check remediation guide with links to the relevant knowledge-base article for every issue found.
Learn more about AI agent readiness →