WebMCP site tools

Written by Andre Molnar 2026-09-08

This is an experiment in implementing WebMCP for this site. It exposes site search and my professional profile as structured, read-only tools that browser-based agents can discover and use.

WebMCP is an evolving browser API rather than a finished web standard. The WebMCP documentation explains the current Chrome implementation, and the draft specification tracks the proposed API.

Try it

You need a recent version of Chrome with WebMCP support and Google's Model Context Tool Inspector extension.

1. Open chrome://flags/#enable-webmcp-testing in Chrome.
2. Set WebMCP for testing to Enabled, then relaunch Chrome.
3. Install the Model Context Tool Inspector extension.
4. Open andremolnar.com and use the inspector to view or invoke the tools registered by the page.

Chrome Canary is only necessary if your regular Chrome installation does not include the testing flag. WebMCP currently requires an open browser tab; these are not remote MCP server tools.

Available tools

search_site

Searches my public projects, blog posts, blurbs and pages. It accepts:

• query — required string between 2 and 200 characters.
• contentTypes — optional array containing blog, blurbs, projects or page.
• tags — optional array of exact tag names. Every supplied tag must match.
• limit — optional integer from 1 to 10; the default is 5.

For example:

{
  "query": "WebMCP",
  "contentTypes": ["projects"],
  "tags": ["WebMCP"],
  "limit": 5
}

An abbreviated response looks like this:

{
  "query": "WebMCP",
  "count": 1,
  "results": [
    {
      "title": "WebMCP site tools",
      "type": "projects",
      "date": "2026-09-09",
      "summary": "An experiment in exposing this static site's search and professional profile as WebMCP tools.",
      "tags": ["WebMCP", "AI", "Web"],
      "url": "https://andremolnar.com/projects/webmcp-site-tools"
    }
  ]
}

get_professional_profile

Returns my public professional summary, strengths, types of work I may be a good fit for, supporting links and contact details. It accepts one optional parameter:

• focus — a string of up to 100 characters used to emphasize an area such as AI, product, design, engineering or leadership.

For example:

{
  "focus": "AI"
}

An abbreviated response looks like this:

{
  "name": "Andre Molnar",
  "location": "Toronto, Ontario, Canada",
  "summary": "Technical leader and hands-on builder with more than 20 years of experience across software engineering, UX, product delivery, and team leadership.",
  "focus": "ai",
  "strengths": [
    "Agent orchestration and workflow design"
  ],
  "goodFitFor": [
    "AI-enabled products and agentic workflows"
  ],
  "evidence": [
    {
      "label": "CV and testimonials",
      "url": "https://andremolnar.com/cv/"
    }
  ],
  "contact": {
    "website": "https://andremolnar.com/",
    "linkedin": "https://www.linkedin.com/in/andretmolnar/"
  }
}