WebMCP: how AI agents use booksmaker.ai
This site exposes its content to AI agents through 12 structured tools. This guide explains what they do and how to try them.
booksmaker.ai supports WebMCP (Web Model Context Protocol), a web standard proposed at the W3C that lets AI agents query a site through named, structured tools instead of parsing raw HTML. On this site an agent can call 12 tools: nine return content (blog articles, pricing, FAQ, genre pages, user guide, release notes) and three act on forms (newsletter, wait-list, blog search).
What is WebMCP?
WebMCP is a set of browser APIs that a website can use to declare its own tools to AI agents running in the same browser. It is incubated at the W3C, co-developed by Google and Microsoft, and available as an early preview in Chromium-based browsers (Chrome 146 and later). Since Chrome 149 an origin trial is running.
The standard defines two ways for a site to expose tools:
- an imperative API, where JavaScript registers tools with names, descriptions and input schemas
- a declarative API, where standard HTML forms carry two extra attributes:
toolnameandtooldescription
booksmaker.ai uses both: nine tools are registered in JavaScript on every page, and three forms on the site (newsletter, wait-list and blog search) are annotated as tools. The official documentation is on developer.chrome.com/docs/ai/webmcp.
Why booksmaker.ai supports WebMCP
People increasingly ask AI assistants instead of browsing. When an assistant can call a tool that returns the pricing table or the answer to a FAQ question, its answer is complete and current. The same tools let an agent fill the newsletter form correctly when you ask it to subscribe you.
WebMCP is a progressive enhancement on this site: on browsers without support, registration is skipped and pages look and work exactly the same. The tools return only public content that is already on the site, and tool usage is not tracked.
The 12 tools available on booksmaker.ai
Nine content tools are registered in JavaScript on every page. Three form tools are attached to the newsletter, wait-list and blog search forms. All content tools accept a lang parameter (it or en) and return content in that language.
| Tool | What it returns | Parameters |
|---|---|---|
listBlogPosts |
Blog articles on AI writing and book creation, filterable by tag | lang, tag, limit |
getBlogPost |
Full content and metadata of one article | slug, lang |
getPricing |
Plans, prices, included features, AI models and their limits | lang |
getFaqAnswer |
FAQ questions and answers matching a keyword | query, lang |
listSolutions |
All genre pages (fantasy, thriller, poetry and more) | lang, filter |
getSolutionPage |
Full details of one genre page | genre, lang |
getHowItWorks |
The step-by-step explanation of how the platform works | lang |
getReleases |
Version history and release notes | lang, version |
getDocs |
User guide sections: dashboard, editor, export, billing and more | lang, section |
subscribeNewsletter |
Fills and submits the newsletter form | email and preferences, via the form |
joinWaitlist |
Fills and submits the wait-list form | name and email, via the form |
searchBlogPosts |
Runs the blog article search | free text, via the form |
The data behind the content tools is the same you see on the pricing page, in the FAQ, in the genre pages, in the user guide and in the release notes.
How to use the tools
WebMCP is still an experimental standard, so trying the tools takes a couple of manual steps:
- Use Chrome 146 or later (or another Chromium-based browser with WebMCP support).
- Enable the experimental flag: type
chrome://flags/#enable-webmcp-testingin the address bar, set the flag to Enabled and relaunch the browser. - Open any page on booksmaker.ai. The tools register automatically on every page of the site.
- Ask an AI agent. An agent running in the browser, or connected to it, sees the list of registered tools with their descriptions and can call them. Try asking: "What pricing plans does this site offer?", "Search the FAQ for exports", "What's new in the latest release?" or "Subscribe me to the newsletter with this address".
To inspect the tools without an agent, Chrome's developer console lists them: open DevTools on any page of the site and call document.modelContext.getTools() to see the twelve registered tools.
Frequently asked questions
Do I need a special browser to use WebMCP on this site?
For now, yes. WebMCP ships as an early preview in Chromium-based browsers and requires the testing flag described above. The standard is still in development at the W3C and support is expected to broaden.
Does WebMCP work with ChatGPT, Claude or Gemini?
Any agent that can operate a WebMCP-capable browser can call the tools, whatever model it runs on. Assistants that answer only from indexed or fetched text cannot call tools yet; for them, the site publishes llms.txt, a plain-text index of its content and tools.
Does using the tools affect my privacy?
No. The content tools only return information that is already public on this site. The form tools submit data only when you ask an agent to fill a form with information you provided. Tool usage is not tracked.
What happens on browsers that do not support WebMCP?
Nothing. Registration is skipped, no error is shown and the site works exactly as before.
Is WebMCP an official standard?
It is a draft incubated in a W3C community group, with an origin trial running in Chrome since version 149 (June 2026). The API may change as the spec evolves, and this page is updated when it does.
Updated on 09/17/2026