How WebMCP Checker Works
A complete guide to using the audit tool and understanding your results.
Enter Your Website URL
Type any URL into the checker input on the homepage and click "Run WebMCP Audit." You don't need to include https:// — just the domain name. You can test any public website: your own site, a competitor's, or a well-known platform to see how it scores.
AI Analyzes Your Site
Our Claude AI engine analyzes the website based on its known technology stack, industry, typical features, and public information. It evaluates the site across 8 technical dimensions relevant to WebMCP readiness. The analysis takes 10–20 seconds.
Get Your Readiness Score
Your score (0–100) appears with one of four verdicts: Agent-Ready (75+), Good Foundation (50–74), Needs Work (25–49), or Not Agent-Ready (below 25). The score reflects how prepared the site's current structure is for WebMCP implementation.
Review the 8 Technical Checks
The Checks tab shows pass/warn/fail status for each of the 8 audit dimensions with a specific finding for your site. Each check explains exactly what was evaluated and why it matters for WebMCP compatibility.
Follow the Fix Recommendations
The Fix It tab provides 4 prioritized recommendations tailored to your specific website type and industry. Each recommendation includes a priority level (High/Medium/Low) and actionable description of what to change.
Copy the Implementation Code
The Code tab provides ready-to-paste WebMCP implementation code tailored to your website's type and use cases. Copy and adapt it for your actual codebase. The Impact tab shows the business case for WebMCP adoption specific to your industry.
Score Guide
The 8 Audit Checks Explained
| Check | What It Tests | Why It Matters for WebMCP |
|---|---|---|
| HTML Form Quality | Whether forms have proper labels, semantic inputs, clear names and types | The Declarative API relies on clean, well-structured HTML forms to register as tools |
| Semantic Markup | Use of appropriate HTML5 semantic elements (nav, main, article, form, button) | Semantic structure helps agents understand page context and tool boundaries |
| JavaScript API Readiness | Whether the site's JS architecture can cleanly integrate navigator.modelContext calls | Tightly coupled SPAs need refactoring; clean service layers are WebMCP-ready |
| Tool Contract Potential | How many clear, callable actions the site's features would translate to as tools | Sites with high tool potential benefit most from WebMCP — more agent use cases |
| Form Action Clarity | Whether form submission actions and outcomes are clear and unambiguous | Agents need to understand what a form does before they can reliably call it as a tool |
| Structured Data / Schema | Existing schema.org or structured data implementation | Sites with structured data have already demonstrated they can think in machine-readable terms |
| Auth Flow Compatibility | Whether authentication flows would work within the browser session model | WebMCP uses browser sessions — complex custom auth flows may need adaptation |
| Agent-Discoverable Actions | Whether the site's key actions would be discoverable via a .well-known/webmcp manifest | Autonomous agents need the manifest to discover what tools your site offers |
Frequently Asked Questions
How long does a WebMCP Checker audit take?
Audits typically complete in 15–30 seconds. The tool fetches your URL, analyzes the site across 8 WebMCP readiness dimensions, and returns a full scored report — usually before you've finished refilling your coffee.
Does WebMCP Checker need access to my backend or code?
No. The audit runs on your publicly accessible URL only. We don't install anything, we don't need credentials, and we don't read server-side code. Everything analyzed is what any browser (or AI agent) would see when it visits your page.
What does an "Agent-Ready" score actually mean?
An Agent-Ready verdict (75–100) means your site has the HTML structure, form semantics, and integration surface an AI agent needs to reliably use your site — whether or not you have formally implemented navigator.modelContext yet. It indicates a site that will become fully WebMCP-enabled with minimal effort.
How is the 0–100 score calculated?
The score is a weighted average of the 8 audit checks. Higher weight is given to checks that matter most for the Declarative API (HTML form quality, semantic markup) and the Imperative API (JavaScript architecture, tool contract potential). A perfect 100 is theoretical — most production sites that are truly agent-ready score 75–95.
Can I re-run the audit after I make changes?
Yes — unlimited free re-runs. We recommend running the audit, picking the highest-impact fixes from the Fix It tab, implementing them, and then re-auditing. Many sites see their score jump 20+ points from the first round of fixes.
Is this a Chrome Canary-based test or a static analysis?
It's a static analysis augmented with the Claude AI API. We don't launch Chrome Canary for each audit — that would take minutes per URL. Instead, the analysis evaluates the same signals an AI agent would extract from your HTML, with AI-generated recommendations tailored to your stack.
What if my score is low — is my site a lost cause?
Definitely not. Most sites score low on their first audit because WebMCP is brand new and no one has implemented it yet. The Fix It tab gives you a prioritized path: start with the top 2–3 recommendations, which typically address 60–80% of the gap. A site going from 35 to 80 in one sprint is common.
Are my audit results saved or shared?
No. We don't persist audit results or the URLs you enter. Each audit is processed in real time and discarded after the report is displayed. There's no database, no cookies tied to audits, and no shared link unless you explicitly generate one on the Compare tool.
Ready to Audit Your Website?
It takes 30 seconds. Enter your URL and get a complete WebMCP readiness report.
⚡ Run Free WebMCP Audit