webmcp-tool

Agent Readiness & WebMCP Scanner

Can AI agents read, operate, and transact on your website?

Audit any public URL across 25 documented checks in seconds: crawler access, server-rendered HTML, form semantics, WebMCP browser tools, and remote MCP discovery — with verifiable evidence and copy-paste fixes.

Free instant score in ~20s. No account, no email required. Works on any public homepage, product, or form URL.

See live sample report (98+/100)Browse 100-site leaderboard

Scans inspect publicly reachable HTTP headers, HTML, discovery endpoints, and scripts. Authenticated or multi-step checkout flows still benefit from a live browser test.

Audit the exact page where customer revenue starts

AI shopping assistants, browser agents, and coding agents land on specific task pages. Enter the product, form, or booking URL where the journey begins.

E-commerce & Catalogs

“Find a 27-inch USB-C monitor under €400 in stock.”

Client-only rendering, missing Product JSON-LD, or blocked AI crawlers make your catalog invisible to shopping agents.

What we verify: SSR product copy, schema.org markup, crawler rules, and agentic payment / WebMCP signals.

E-commerce agent readiness guide →

B2B Lead & Support Forms

“Request a quote with my company details and requirements.”

Unlabelled inputs, custom div dropdowns, and CAPTCHA gates stop browser agents right before conversion.

What we verify: Explicit labels, control names, autocomplete tokens, error semantics, and CAPTCHA friction.

Forms agents can complete →

SaaS & Booking Flows

“Check availability on Friday afternoon and prepare a booking.”

Opaque JS widgets without declarative tool contracts force agents to guess clicks instead of calling structured tools.

What we verify: Declarative form tool annotations, document.modelContext tools, and /.well-known/mcp.json discovery.

Declarative WebMCP guide →

From score to shipped fix in one afternoon

Every public scan produces a shareable URL with your Fitness vs. Adoption split, the top three highest-gain fixes, and an implementation brief for your engineering team or coding agent.

01

Separate Fitness from Adoption

See immediately whether agents fail on basic HTML/form fundamentals (65 pts) or missing WebMCP / MCP protocols (35 pts + 15 bonus).

02

Verify every finding

Inspect the exact HTTP header, robots.txt rule, form control count, or discovery path observed during the crawl.

03

Ship & re-scan

Copy tailored WebMCP tool definitions, manifest JSON, and llms.txt templates — or unlock all 25 checks by email.

See live sample report (98+/100) →

Implementation & Advisory

Want us to make your product agent-ready or build your MCP server?

Whether you need high-impact HTML/form fixes, browser WebMCP tools, or a production OAuth 2.0 remote MCP server prepared for Anthropic and OpenAI directory review — talk directly with Gorden.

Book a 30-min technical callRequest a fixed-price quote

For SEO, GEO & Web Agencies

Turn Agent Readiness into a billable client audit

Benchmark a prospect against two competitors in 60 seconds, share side-by-side /compare links, and hand their developers a prioritized 25-check engineering roadmap.

Read the agency audit playbook →

Transparent scoring methodology — zero black-box grades

Every rule, point weight, and conditional N/A exclusion is published openly. Non-applicable checks (such as payment rails on a non-commerce site) are removed from the denominator instead of counted against you.

25 checks · 6 pillars · 100 base points (+15 bonus)

Agent Access

20 base points

4 checks

Machine Readability

20 base points

5 checks

Actionability

25 base points

5 checks

Integration

15 base points

3 checks

Identity & Commerce

10 base points

3 checks

WebMCP Implementation

10 base points (+15 bonus)

5 checks

WebMCP includes 10 base points plus up to 15 depth bonus points (score capped at 100).

Read the full 25-check methodology →What is WebMCP? →

WebMCP Tool

One ecosystem: Audit, Build, and Measure AI Agents

WebMCP Tool audits public agent readiness. Our sister service mcp-tool.com builds and hosts production remote MCP servers at a fixed price. agenttracking.co measures real AI crawler and agent visits on your domain.

Fixed-price MCP servers (mcp-tool.com) →Measure agent traffic (agenttracking.co) →Call this scanner via MCP →

Frequently asked questions

What does the Agent Readiness scan actually test?

It fetches your live URL and evaluates 25 documented checks across 6 pillars: Crawler & Agent Access (18 pts), Machine Readability (22 pts), Actionability (25 pts), WebMCP (10 base + 15 bonus pts), Agent Discovery (15 pts), and Interoperability & Commerce (10 pts).

Is the scan free without creating an account?

Yes. Your overall score, Fitness vs. Adoption breakdown, pillar bars, top three fixes, and tailored WebMCP starter snippets are immediately public and shareable without signup. Unlocking the full 25-check line-by-line report only asks for an email address.

What is the difference between WebMCP and Anthropic's MCP?

WebMCP (document.modelContext) runs inside the user's browser tab so in-page AI assistants can invoke your forms and UI actions using the user's existing session. Remote MCP (/.well-known/mcp.json) exposes an HTTP JSON-RPC endpoint for headless agents like Claude, ChatGPT, and Cursor. Strong sites support both.

Which URL should I scan — my homepage or a subpage?

Start with your homepage to verify site-wide discovery files (robots.txt, llms.txt, /.well-known/mcp.json), then scan the specific product, pricing, contact, or booking subpage where conversions happen.

See how AI agents experience your website today

Run a live 25-check audit in ~20 seconds — free, public, and backed by raw evidence.

Free instant score in ~20s. No account, no email required. Works on any public homepage, product, or form URL.