Add one script tag. Define what AI assistants can do. Watch them search, click, and transact — on behalf of your users.
<!-- Add WebMCP to your site -->
<script
src="https://aigentably.com/mcp-manager.js"
data-site="your-site-id"
></script>How it works
WebMCP is an emerging open standard that lets AI assistants discover and call tools on any website — the same way HTTP lets browsers load pages. Claude, ChatGPT, and Cursor already speak it.
Aigentably handles all the infrastructure: hosting your tool manifest, serving it at the right URL, and giving you a dashboard to manage everything — no backend required.
Learn more about WebMCPAI agent visits your site
Detects the Aigentably script and fetches your tool manifest
Reads your available tools
Sees what actions are defined — addToCart, search, submitForm…
Calls the right tool
Executes the JavaScript in the user's browser, on your real site
Returns the result
The agent gets structured output and responds naturally to the user
See it in action
Once your site has WebMCP tools, any AI assistant can take real actions — no scraping, no custom API.
Works wherever your users are
One script tag. No build step required. Your tools are discoverable by AI agents in under 30 seconds.
Paste your URL and AI suggests the right tools for your site — add to cart, search, contact forms, and more.
See which AI agents are calling your tools, success rates, latency, and caller breakdown by agent type.
Show the right tools on the right pages. Scope any tool to a specific URL path — product pages, checkout, blog, and more.
Run any tool on your real site with one click. See the result instantly — no AI agent required to test your setup.
Enable or disable tools individually. Live security linting warns you about dangerous patterns before you save.
Copy one line of code into your HTML. No build steps, no dependencies.
Describe what actions AI agents can take on your site using simple JSON schemas or pick from 42 pre-built templates.
See real-time data on how AI agents interact with your product, success rates, and caller breakdown.
{
name: "addToCart",
description: "Add product to cart",
inputSchema: {
productId: { type: "string" },
quantity: { type: "number" }
}
}Start free. Get in touch for Pro.
For side projects and testing
For production applications
We'd love to hear from you. Reach out and we'll get back to you shortly.
Join the first wave of developers building for the agent-first web.
Get started for free