WEBMCP.COM

The WebMCP Directory.

A live directory of websites exposing WebMCP tools to AI agents. Browse the tools each site registers, inspect their input schemas, or query the agent-readable JSON API. Built on the W3C webmachinelearning/webmcp proposal and maintained by nekuda.

Browse websites agents can use.

6,656 tools indexed
tool mix Answerread-only: search, details, availability. No side effects. Actiondrives the page: carts, forms, navigation. Reversible. Sensitive Actionmoney or commitment: checkout, booking, subscribing. Read the full methodology →
Answer 55%Action 40%Sensitive Action 5%

Frequently Asked

What is WebMCP?

WebMCP lets websites give AI agents a direct way to interact with them. Instead of trying to click buttons and figure out a page like a human, AI can use actions the website provides.

There are two main ways people use WebMCP. The first is co-browsing, where you interact with AI built into your browser, such as Gemini in Chrome. The second is remote browsing, where AI applications like ChatGPT or Claude open a browser and complete tasks on your behalf, with or without you watching. Both ChatGPT and Claude have recently added browser support.

Because websites tell AI exactly what they can do, interactions become faster, more reliable, and more predictable than screen scraping. webmcp.com is a live directory of websites that support WebMCP.

How is WebMCP different from MCP?

MCP connects AI applications to servers and services. You can think of it as an API that lets AI access capabilities outside the application itself.

WebMCP brings that idea into the browser. The stated goal of the WebMCP proposal is to enable co-browsing, where users stay on a website and interact with it alongside their AI assistant, rather than the AI operating in a separate application.

In many cases, this is better for everyone. Users can navigate complex interfaces more easily while staying in control, and website owners keep users engaged on their own site instead of moving the experience into a headless browser.

WebMCP can also be used from remote automated browsers. However, website owners can choose to block remote automated browser access using existing bot protection or other access controls.

How do I get my site listed?

Ship WebMCP tools on your site, then submit it through the “Add your site” box on webmcp.com. Our scanner verifies the tools are live before listing.

Can AI agents and crawlers read this directory programmatically?

Yes. Curated directory records are available as JSON at /api/v1/sites and /api/v1/sites/<host>. Shopify stores have a separate index: read shared tools at /api/v1/platforms/shopify and search stores at /api/v1/platforms/shopify/stores?q=. The OpenAPI spec is at /api/openapi.json. See the API for Agents page.

For Agents

API for Agents Full API docs

Query the directory programmatically — list sites, inspect each tool's input schema, probe a URL with /api/v1/lookup, or pull aggregate counts. Read-only JSON, no auth, CORS open. Stub-only sites (modelContext present but no registered tools) are excluded.

GET /api/v1/lookup?url=… Does this URL expose WebMCP? GET /api/v1/sites?type=live List live WebMCP sites GET /api/v1/sites/{host} One site's full capabilities GET /api/v1/stats Directory-wide counts + top sites GET /api/v1/platforms/shopify Shared Shopify tools and last check GET /api/v1/platforms/shopify/stores?q=… Find Shopify stores separately

OpenAPI 3.1 spec at /api/openapi.json · full reference at /api-docs