A scanner for the agentic web, built by the team closing its biggest gap

What AX Audit is

AX Audit scores any domain against the open AgentReady (AR-*) standard — five categories that describe how an autonomous agent moves through the web: finding a site, understanding it, calling its capabilities, authenticating against it, and paying it.

Every check maps to a real, cited piece of evidence — the actual header, file, or response an agent would encounter — plus a fix prompt you can paste straight into your coding agent. No login, no credit card, one free report per scan.

Why AI readiness matters now

Search once ran on the assumption that your audience was a person with a browser. That assumption is breaking. A growing share of traffic to any given site is now an agent — a coding assistant checking your docs, a shopping agent comparing your pricing, a research agent reading your terms — and most sites were never built for that reader. They bury pricing in JavaScript a crawler can't execute, skip llms.txt and machine-readable terms entirely, and have no path for an agent to do anything but read.

Being agent-ready is becoming table stakes the same way mobile-responsive and HTTPS once were. The sites that measure it first fix it first — and the rungs above basic discovery (authentication, payment, and especially provisioning) are where almost nobody has built anything yet. AX Audit exists to make that gap visible, in plain evidence, in under a minute.

Who built it

AX Audit is built and maintained by AINative Studio, which builds the infrastructure that makes agent-readiness actionable rather than theoretical: ZeroDB for agent memory, Agent Cloud / Swarm for running agents at scale, and an MCP registry for discoverable tools. AX Audit is also our own reference implementation — we score ourselves against the same six categories we check for you, in public.