Agent Readiness Score: The Free Tool That Analyzes Whether Your Website Is Visible to AI Agents

TL;DR

The Agent Readiness Score is a free GEO Metrics tool that analyzes any domain and returns a technical score from 0 to 100 on how readable, accessible and recommendable your website is to AI agents. It evaluates 4 dimensions: Discoverability, Content Accessibility, Bot Access Control and Protocol Discovery, plus GEO and AI Citation Signals. No sign-up. Results in seconds. It is the complementary tool to the GEO Content Readiness Score: one analyzes your website technically, the other analyzes your content. → Analyze my website for free

In 2026, web traffic no longer comes only from people clicking links. It also comes from AI agents that crawl, read and interpret your site to decide whether your brand deserves to be recommended.

The problem: most websites are built for browsers and for Google. Not for AI agents. And the technical difference between a site that is readable by an agent and one that isn't can determine whether your brand appears in ChatGPT, Perplexity or Claude responses — or doesn't.

The Agent Readiness Score by GEO Metrics measures exactly that difference — for free, no sign-up, in seconds.

What Is the Agent Readiness Score

The Agent Readiness Score is a free GEO Metrics tool that analyzes the technical infrastructure of any website and determines whether it is correctly configured for AI agents to read, understand and recommend it.

Unlike the GEO Content Readiness Score — which analyzes whether the content of a URL is structured to be cited — the Agent Readiness Score analyzes the full domain: robots.txt, sitemap, llms.txt, HTTP headers, structured data, agent authentication protocols and authority signals.

It is the technical layer. The Content Readiness Score is the editorial layer. They are complementary tools, not substitutes.

In seconds, the tool delivers:

• A global score from 0 to 100 with a readiness level

• 4 analysis dimensions with individual scores

• Per sub-metric scoring within each dimension

• Status for each check (✓ correct / ⚠ improvable / ✗ missing)

• Exact technical instructions for implementing each improvement

• Full report download as PDF

No forms. No account. Just the domain to analyze.

The 4 Dimensions the Score Evaluates

1. Discoverability (up to 100 points)

Measures whether AI agents can find and navigate your site efficiently.

Sub-checks:

• robots.txt — Does it exist and have clear rules? Does it include references to sitemaps?

• Sitemap — Is it valid and accessible at the standard path?

• HTTP Link Headers — Do they include agent-useful relation types such as rel="api-catalog", rel="service-desc" or rel="describedby"?

A site with a correct robots.txt and sitemap but no agent-oriented Link headers can score 84/100 on this dimension — solid on the basics, with room to improve on the advanced layer.

2. Content Accessibility (up to 100 points)

Measures whether your site's content is readable by language models under optimal conditions.

Sub-checks:

• llms.txt — Does the file /llms.txt exist with useful content for LLMs? And /llms-full.txt for complete coverage?

• Markdown Negotiation — Does your server return content in Markdown format when an agent requests it with Accept: text/markdown? If it always returns HTML, models have to parse it — far less efficient than receiving clean text

Markdown negotiation is the highest-impact improvement in this dimension and one of the least implemented in the current web ecosystem.

3. Bot Access Control (up to 100 points)

Measures whether your site has explicit, clear policies on how AI crawlers should access your content.

Sub-checks:

• AI bot rules — Are there explicit rules for the main AI crawlers (GPTBot, ClaudeBot, ChatGPT-User, PerplexityBot, GoogleExtended)?

• AI bot access — Do the 26 main AI bots have access defined explicitly, or does the wildcard User-agent: * leave intent unclear?

• Content signals — Does robots.txt include Content-Signal directives declaring your preferences on AI training use of your content?

• Web bot authentication — Is there a JWKS directory at /.well-known/http-message-signatures-directory so your site can cryptographically identify itself to agents?

This dimension is the least understood by most development teams — and one of the most relevant in the context of 2026 AI regulation.

4. Protocol Discovery (up to 100 points)

Measures whether your site has the advanced agentic infrastructure that allows AI agents to interact with it autonomously.

Sub-checks:

• MCP Server Card — Does a Server Card exist at /.well-known/mcp/server-card.json with MCP server info, transport endpoint and capabilities?

• Agent Skills Index — Is there a skills index at /.well-known/agent-skills/index.json with the actions your site can execute?

• WebMCP — Does your page implement the WebMCP API with navigator.modelContext.provideContext() and tool definitions?

• API Catalog — Is there an API catalog at /.well-known/api-catalog with OpenAPI specifications?

• OAuth / OIDC Discovery — Does your site publish authentication metadata at /.well-known/openid-configuration?

• OAuth Protected Resource — Does the file /.well-known/oauth-protected-resource exist with authorization servers and supported scopes?

This dimension is the boundary between a traditional website and one prepared for the agentic web. Most sites score 0/100 here in 2026 — not because they are poorly built, but because these standards are recent and adoption is in its early stages.

5. GEO and AI Citation Signals (up to 100 points)

Measures whether your site has the signals that lead language models to trust your content as a citable source.

Sub-checks:

• Structured data — Is there JSON-LD with types such as WebSite, Organization, LocalBusiness, Article?

• Meta tag completeness — Are meta description, og:title, og:description and og:image all present?

• Authorship and authority signals — Does the content include rel='author', JSON-LD with author/publisher fields, an /about page with organization information?

A Real Example: Analysis of leadscoring.ai

To illustrate how the tool works, here are the results from analyzing a real site processed with the Agent Readiness Score:

The analysis reveals a common pattern in 2026: the basics are well implemented (robots.txt, sitemap, structured data, meta tags), but the agentic layer is completely absent. A Protocol Discovery score of 0/100 does not mean the site is broken — it means it is not prepared for the next generation of AI agents that are already operating in production.

Every red check comes with the exact technical instruction for implementing the fix — including the file path, content format and, in many cases, the code snippet to implement.

The Difference Between the Agent Readiness Score and the GEO Content Readiness Score

Both tools are free, both return a score from 0 to 100 and both generate concrete actions. But they analyze different layers:

The correct sequence is to use the Agent Readiness Score first — to make sure the technical infrastructure is in order — and then the Content Readiness Score to optimize each piece of content.

Why Agentic Infrastructure Matters in 2026

AI agents are not just chatbots that answer questions. They are autonomous systems that crawl the web, evaluate sites, execute tasks and make decisions about which sources are reliable and which are not.

In that context, a website without llms.txt, without explicit rules for AI bots and without an MCP Server Card is the equivalent of a site without a sitemap in 2015 — functionally visible to humans, but invisible to the infrastructure that increasingly determines traffic and influence.

The good news: most of these improvements are targeted technical implementations, not redesigns. An llms.txt file, correct HTTP headers and explicit bot rules in robots.txt can raise the score 20-30 points in a single working session.

How to Use the Agent Readiness Score

1. Go to trygeometrics.com/agent-readiness-score

2. Enter your website's domain

3. The tool analyzes in seconds and delivers the full report

4. Review the red checks — each includes the exact implementation instruction

5. Implement the improvements and re-analyze to verify progress

6. Download the PDF to share with your technical team or client

Frequently Asked Questions About the Agent Readiness Score

Is it completely free? Yes. The Agent Readiness Score is 100% free and requires no account or payment information. You only need the domain of the site you want to analyze.

What is the difference from the GEO Content Readiness Score? The Content Readiness Score analyzes the editorial content of a specific URL — how well written and structured it is for AIs to cite it. The Agent Readiness Score analyzes the technical infrastructure of the full domain — whether AI agents can access, read and process the site. They are complementary: one optimizes content, the other optimizes infrastructure.

What is llms.txt and why does it matter? llms.txt is an emerging standard file published at the root of the domain (/llms.txt) that provides language models with a structured summary of the site, its main sections and its most relevant URLs. It is the robots.txt equivalent for the AI era — it tells LLMs what they will find on the site and how to navigate it efficiently.

What is the MCP Server Card and who needs to implement it? The MCP Server Card is a JSON file at /.well-known/mcp/server-card.json that describes your MCP server's capabilities: name, version, transport endpoint and available tools. It is relevant for sites that want AI agents to interact with their APIs or features autonomously. In 2026, it is an advanced implementation — but the adoption curve is accelerating.

Does a low score mean my site is poorly built? No. A low score on Protocol Discovery, for example, simply means the site has not implemented emerging agentic standards — not that it has performance, security or user experience issues. The Agent Readiness Score measures specifically the readiness for the AI agent ecosystem, not the overall quality of the site.

How long does it take to implement the improvements? It depends on the starting score. The highest-impact improvements (llms.txt, Link headers, explicit bot rules in robots.txt) can be implemented in 1-2 hours by a developer with server access. Advanced improvements (MCP Server Card, Agent Skills, OAuth) require more time but are optional for most sites in 2026.

Can I use it to analyze client sites? Yes. The tool is public and requires no authentication. You can analyze any public domain. For agencies, it is a fast way to generate a technical AI visibility audit for a prospect before a meeting.

Analyze your domain now. It's free, takes seconds, and the result tells you exactly what to implement.

Analyze my website for free → trygeometrics.com/agent-readiness-score

Want to also know whether your content is structured to be cited by AIs? Use the GEO Content Readiness Score → trygeometrics.com/geo-readiness-score

Want to see how your brand appears in ChatGPT, Gemini, and Perplexity right now? Request a demo at trygeometrics.com.