How to Enable WebMCP in Chrome 146: Step-by-Step Guide

WebMCP launched in early preview in Chrome 146 in February 2026. It is available now — but you need to enable a feature flag and install a debugging extension to get started. This guide walks you through the full setup in under 10 minutes.

What you will need

• Chrome 146+ (Canary or Beta channel recommended)
• 5-10 minutes
• No coding required for basic setup
• The Model Context Tool Inspector extension (free, from Chrome Web Store)

Step 1: Install the Right Chrome Version

The stable Chrome release does not yet include WebMCP. You need Chrome 146 or higher from the Canary or Beta channel.

Option A: Chrome Canary (Recommended for Developers)
Chrome Canary is the most cutting-edge build and always has the latest features. It can run alongside your regular Chrome installation without conflicts.

1. Go to google.com/chrome/canary/
2. Download and install Chrome Canary
3. It installs separately — your regular Chrome is unaffected

Option B: Chrome Beta
More stable than Canary but still includes the WebMCP flag. Go to google.com/chrome/beta/ and follow the same steps.

Version check
Open Chrome menu → Help → About Google Chrome. You need version 146.0.7672.0 or higher. If your version is lower, update before proceeding.

Step 2: Enable the WebMCP Flag

Chrome feature flags are experimental settings that activate features before they are officially released.

1. Open Chrome Canary or Beta
2. In the address bar, type: chrome://flags and press Enter
3. In the search box at the top, type: WebMCP
4. You will see the flag “WebMCP Testing” — click the dropdown next to it
5. Select “Enabled”
6. Click the blue “Relaunch” button at the bottom of the screen

// Direct URL to the flag (paste into Chrome address bar):
chrome://flags/#enable-webmcp-testing

After relaunching
Chrome restores your previous tabs. The WebMCP API is now active — navigator.modelContext will be available on HTTPS pages.

To verify: open any HTTPS site, open DevTools (F12), go to Console, and type: “modelContext” in navigator
If it returns true, WebMCP is enabled correctly.

Step 3: Install the Model Context Tool Inspector

The Inspector is a Chrome DevTools extension that shows you all WebMCP tools registered on the current page — and lets you test them with custom inputs.

1. Open the Chrome Web Store (chromewebstore.google.com)
2. Search for “Model Context Tool Inspector”
3. Click “Add to Chrome” and confirm the installation
4. The extension adds a new tab to Chrome DevTools

To access it:

1. Open DevTools: press F12, or right-click any page → Inspect
2. Click the “>>” overflow tab if the Inspector tab is not visible
3. Select “Model Context Tool Inspector”
4. Navigate to a page with WebMCP tools — they will appear in the panel

Step 4: Test on Google’s Live Demo

Google provides a live flight search demo that showcases both the Declarative and Imperative APIs. It is the fastest way to see WebMCP in action.

1. With WebMCP enabled, navigate to the Google Chrome AI flight demo (linked from developer.chrome.com/blog/webmcp-epp)
2. Open the Model Context Tool Inspector
3. You will see tools like searchFlights, listFlights, and setFilters registered on the page
4. Click a tool, enter test parameters (e.g., origin: “JFK”, destination: “LAX”, date: “2026-06-15″)
5. Click Execute and observe the structured JSON response

This shows exactly what an AI agent sees when it interacts with a WebMCP-enabled site. Your tools will appear the same way once implemented.

Step 5: Test Your Own Site

Once you have implemented WebMCP tools on your site, use this workflow to verify them:

1. Navigate to your site in Chrome with the flag enabled
2. Open DevTools → Model Context Tool Inspector
3. Confirm your tools appear with correct names and descriptions
4. Test each tool with realistic parameters
5. For write tools (readOnly: false), confirm the browser confirmation prompt appears
6. Check the browser console for any JavaScript errors

Troubleshooting

What’s Next

With Chrome configured and the Inspector installed, you are ready to implement and test WebMCP tools on your own site. Recommended reading order:

1. How to Make Your Website Agent-Ready (start here for the full process)
2. WebMCP Declarative API Guide (for HTML form annotation)
3. WebMCP Imperative API Guide (for custom JavaScript tools)
4. WebMCP Security Best Practices (before going live)

Ready to implement WebMCP?

Rather have experts handle the implementation? We set up, test, and document WebMCP tools for your site — so you do not have to navigate the early preview documentation yourself.

If you want your own site to work with AI agents, see our WebMCP implementation service, or read the background in WebMCP vs MCP.

WebMCP Audit Now

Related Reading

• WebMCP Declarative API Guide
• WebMCP Imperative API Guide
• WebMCP Implementation Services