# Is your site ready for AI agents? A free 10-second agent-readiness check

> AI agents already browse sites instead of humans. We break down the 6 machine-readability signals — llms.txt, structured data, author markup, sameAs, markdown alternates — and how to push your agent-readiness score to 100.

_Source: https://seomatrix.ai/blog/agent-readiness-checker/ · Updated: 2026-07-03_

---

Google added an **Agentic Browsing** audit to Lighthouse, Cloudflare shows sites an **Agent Readiness score**, and BrightEdge predicts that by the end of 2026 most "visitors" to commercial sites will be AI agents. "[Can an agent consume my site?](/tools/ai-view/)" has become a discipline of its own — so we built a free checker for it: the [**AI Agent-Readiness Checker**](/tools/agent-readiness/).

## What the checker scores

Paste a domain — get a **0–100** score across six machine-readability signals (weights in brackets):

1. **Structured data, JSON-LD (30)** — the only format where an agent gets facts about the page without parsing prose: types, prices, authors, FAQs.
2. **llms.txt (20)** — a curated site index for AI engines (the [llmstxt.org standard](/blog/llms-txt/)): what matters, where it lives, what to read first.
3. **Author / Person markup (15)** — the E-E-A-T anchor: who stands behind the content. A standalone Person node with `knowsAbout` and `worksFor`, not just a name in the footer.
4. **sameAs entity links (15)** — connecting your brand to the knowledge graph (social profiles, Wikidata) so an agent knows WHICH entity you are.
5. **llms-full.txt (10)** — your whole site's content in one fetch, so an agent doesn't crawl a hundred pages.
6. **Markdown alternates (10)** — `rel="alternate" type="text/markdown"`: a clean markdown version of the page instead of tons of HTML/JS — the format RAG pipelines love.

A score of 75+ is "agent-ready", 50–74 "partially ready", below that agents see your site as a brick wall.

## We tested it on ourselves — honestly

We ran the checker on our own homepage and scored **75/100**: no standalone Person node (the author markup was hiding nested inside Organization) and no markdown alternates for the homepage. Cobbler's children. We fixed both the same day — [seomatrix.ai](https://seomatrix.ai/) now serves `/index.md` and a proper author node. Check us — and yourself.

> Every failing signal maps to a concrete fix: for markup there's the [Author/E-E-A-T schema generator](/tools/author-schema/) and the [Schema.org generator](/tools/schema-generator/); for the index — the [llms.txt generator](/tools/llms-txt-generator/) and the [validator](/tools/llms-validator/).

## When a score isn't enough

The checker answers "can agents read us". The next question — "does AI actually cite us on real queries?" — is covered by the [free AI-visibility check](/ai-visibility/), and the full picture with an element-level prioritized fix list comes from the [paid GEO/SEO audit](/audit/) — which now includes a free month of AI-visibility monitoring.

[**→ Check your site's agent-readiness**](/tools/agent-readiness/)

## FAQ

### What is site agent-readiness?

It's a site's ability to be CONSUMED by an AI agent or answer engine — not "render nicely in a browser" but expose machine-readable structure. The key signals are llms.txt, llms-full.txt, JSON-LD, author markup, sameAs entity links and markdown page versions.

### How is agent-readiness different from classic SEO?

Classic SEO optimizes for a crawler and blue-link rankings. Agent-readiness optimizes for an agent that reads your site as DATA — fetching, citing and acting on it. A site can rank perfectly and still be unreadable to agents.

### How do I check agent-readiness for free?

Paste your domain into our free checker at /tools/agent-readiness/ — you get a 0–100 score across six signals, and every failing signal explains what to fix. No signup.