AI agent readiness

Check whether agents can find and read a clean machine-readable version of every public page. The report tests the content-site contract directly, then keeps unrelated API, application and commerce checks out of the failure count.

Install and run the audit

Install the CLI once, then run seo start to connect your site and Google data. After setup you can run this audit yourself or ask your agent to run it.

Runafter install

Install the SEO CLI globally.

npm i -g seo

Run setup after installation.

seo start

Check the public content-site contract with a fresh bounded crawl.

seo agent-readiness https://example.com --max-pages 100 --json

What you get from this audit

Check whether agents can find, fetch, and read a stable machine-readable version of every public content page.

Best for: A documentation, publishing, or content site serves machine-readable versions of its pages.
Needs: Fresh content-site crawl
You get: An unscored content-profile result with pass, warning, fail, unknown, information, and not-applicable states kept separate.
Use next: Check AI search technical readiness

What you need before you run it

This audit uses public HTML pages, Markdown alternatives, HTTP content negotiation, route manifests, Agent Skills, llms.txt, crawler policies, and structured identity.

Fresh content-site crawl

Provides the public HTML route inventory, response evidence, crawler policy, and structured identity found during this run.

Agent-facing representations and discovery

Provides Markdown alternatives, negotiated responses, the route manifest, Agent Skills, and llms.txt for direct validation.

What the crawl evidence checks

Every finding stays tied to the page, rule and crawl scope that produced it. Check what the crawler reached before treating an issue count as complete.

Confirms that every successful public HTML page advertises one working Markdown alternative with a matching canonical relationship.

Compares explicit and negotiated Markdown byte for byte, honours Accept q-values, and checks repeated response hashes for stability.

Validates Agent Skills digests, llms.txt links, route-manifest coverage, crawler access, protocol handling, and the site identity graph.

Read the coverage before the findings

A clean result means little if important pages were blocked, skipped or outside the crawl limit. Coverage comes first, then findings.

Read first	Why it matters
Result	An unscored content-profile result with pass, warning, fail, unknown, information, and not-applicable states kept separate.
More detail	Affected URLs, observed evidence, and a concrete action for every representation or discovery check that needs attention.
Read it with	Start with failed routes and mismatched bytes. A missing optional file can be useful context, but it is not a search defect or proof that agents cannot use the site.

What the result cannot prove

This report has no aggregate score. A clean result cannot prove that an AI service fetched, used, mentioned, cited, ranked, or selected a page.

Start with failed routes and mismatched bytes. A missing optional file can be useful context, but it is not a search defect or proof that agents cannot use the site. Repeat the same public route scope after deployment and compare the affected checks.

This is a useful post-deploy check for documentation and content sites. Ask your agent to repeat the same route scope and report new failures, resolved warnings, and any route that disappeared from the HTML or Markdown inventory.

Use a different report for these jobs

You need an AI mention, citation, referral, ranking, or selection measurement. The main job is testing a public API, interactive agent application, or commerce flow.

You want to check technical eligibility for Google AI search features rather than the machine-readable content contract.

Run AI search readiness. It checks crawl, index, canonical, snippet, and page evidence without treating optional agent discovery as a search requirement.

AI search readiness audit

You need observed mentions, citations, prompt coverage, or share of voice inside AI answers.

Use a repeatable external visibility measurement. This audit can prove whether the site delivered its content contract, but it cannot show whether an AI product used it.

AI visibility tracking

Use the audit with an agent or in code

The command above is enough for most work. These options let an agent, script or application run the same audit with explicit input and structured output.

Use the report catalog when an agent, script or CI job needs predictable JSON instead of the normal terminal output. Check the current input first, then run the audit with the validated parameters.

Check the current input before the agent or script runs the audit.

seo reports describe agent-readiness --json

Once the input is valid, run the audit through the report catalog.

seo reports run agent-readiness --params '{"url":"https://example.com/","maxPages":100}' --json

An MCP agent follows the same two-step flow. It inspects the current schema, fills the required fields and only then runs the audit.

Ask the MCP server for the current audit input.

{
  "id": "agent-readiness"
}

Run the audit after the agent has checked and filled the input.

{
  "id": "agent-readiness",
  "params": {
    "url": "https://example.com/",
    "maxPages": 100
  }
}

Install the package in a Node 22 project when your application needs to run the audit directly. The result includes the same evidence, warnings and limits returned through the CLI and MCP server.

Add the SEO package to your project.

npm install seo

Run the audit from your TypeScript application.

import { executeReport } from 'seo/mcp'

const result = await executeReport(
  'agent-readiness',
  {
  "url": "https://example.com/",
  "maxPages": 100
},
)

console.log(result)

What to do next

Fix shared representation failures before patching one generated file. Repeat the same route scope after deployment and confirm the explicit Markdown URL, negotiated response, discovery files, and identity evidence now agree.

01Fix shared generation or delivery problems before patching individual Markdown files.
02Repeat the same public route scope after deployment and compare the affected checks.

Review crawl access, indexability, snippet controls, page structure, and optional agent resources without inventing an AI visibility score.

AI search readiness audit

Check whether an optional llms.txt file is reachable, readable, and consistent with useful pages from the crawl.

llms.txt audit

Review naming, authorship, dates, schema, and sameAs links without claiming a search engine has recognised an entity.

Entity SEO audit

Find technical SEO issues across a site and save the crawl for follow-up audits.

Technical SEO site crawl audit