Best fit
Website owners and small teams
- People who got a low score from a checker and need a roadmap
- Teams that can edit HTML and want a clean first implementation
- Operators who need a practical path before paying for a custom audit
Pack #17 · WebMCP Agent-Ready Website Toolkit
A complete DIY toolkit to make your website agent-ready: WebMCP annotation templates, an agent-readiness audit checklist, a tool description copywriting guide, and Chrome 149 testing instructions. Prepare for the agent-ready web — not certified agent-ready.
Prepare for the agent-ready web — without $200-$500 agency fees.
Who it's for
The toolkit is aimed at website owners, small teams, and developers who want the agent-readable surface without paying for a broad custom engagement before they know what needs changing.
Best fit
Also useful for
Not a fit
Prepare for the agent-ready web with the current WebMCP-era surface, not a certification claim.
WebMCP tools
Use the annotation templates to make contact, booking, product search, and support flows discoverable to agents without rebuilding the entire site.
Readable context
The audit checklist covers llm.txt, schema, robots.txt, and the visible copy around the tool so agents know what the site does before they call it.
Chrome 149 testing
The testing instructions walk through Chrome 149 setup, discovery checks, prompt-based verification, and the common mistakes that break tool use.
An agent-ready website is a site that exposes useful actions and supporting context in a way AI agents can discover and use: WebMCP tools, readable labels, schema, llm.txt, and a testing loop that verifies the result. It is not a certification or a permanent-compatibility promise — WebMCP is still an origin trial, and the API surface may change. Being agent-ready means you have done the work to let agents like ChatGPT, Claude, and Gemini find your site's functionality and invoke it on behalf of their users.
The practical version: agents can discover what your site does (via llm.txt and schema), understand what actions are available (via readable labels and tool descriptions), and actually take those actions (via WebMCP tools). You have tested the loop with the inspector and confirmed it works. That is the bar.
Everything in the pack is there to help you get from audit to implementation without wandering.
Audit
A practical review path from "I ran the checker" to "I know what to fix next." Use it to decide what to annotate, what to rewrite, and what to leave alone.
Templates
Copy-paste HTML annotation templates for the common website actions the pack targets, with the boring bits explained so the markup stays editable.
Writing
Rules and examples for writing tool descriptions that agents can discover and invoke instead of skipping over vague or overstuffed labels.
Testing
Step-by-step instructions for enabling the current trial, checking discovery, and confirming that the tool behaves the way the template says it should.
Launch
A scene-by-scene screencast script you can use to ship the pack, train a teammate, or turn the workflow into future sales material.
Pricing ladder
The page positions the toolkit as a prep product first, with audit and implementation services for teams that want someone to apply the workflow for them.
Digital pack
The DIY toolkit itself: templates, checklist, copywriting guide, and Chrome testing instructions.
Implementation audit
A paid review that tells you what to annotate, what to rewrite, and what to test first.
Full implementation
Hands-on implementation for teams that want the toolkit applied for them instead of doing it themselves.
Answers to the search questions this landing page is built to catch.
FAQ
A site that exposes useful actions and supporting context in a way AI agents can discover and use: WebMCP tools, readable labels, schema, llm.txt, and a testing loop that verifies the result.
FAQ
No. This page and toolkit are preparation for the agent-ready web, not a certification or a permanent-compatibility promise. WebMCP is still evolving.
FAQ
Not for the first path. The pack is designed to start with declarative HTML templates. The imperative JavaScript examples are for sites that need richer interactions.
FAQ
Because a checker tells you whether your site is ready. It does not give you the template files, the tool-description wording, or the testing instructions needed to actually fix it.
FAQ
How to enable the current trial, verify discovery, exercise the inspector, and catch description or schema mismatches before you publish.
FAQ
You still keep the workflow: audit, annotate, write better descriptions, and test again. The current pack is aligned to the origin-trial-era surface, so it is meant to be adapted as the standard evolves.
Next step
Start with the digital pack if you want the roadmap and the templates. Start with the audit if you want a second pair of eyes. Start with full implementation if you want the work done for you.
Prepare for the agent-ready web — this page is a roadmap, not a certification.
Adjacent field notes that help frame the implementation path and security posture.
Field note
A step-by-step walkthrough of declarative and imperative WebMCP tool registration, testing with the inspector, and Chrome 149 origin trial setup.
Read more →Field note
How to write, validate, and deploy llm.txt for AI agent discoverability — with examples, a comparison to llms.txt, and a checklist.
Read more →Field note
Why the testing loop matters once agents start touching real web tools and browser workflows.
Read more →Field note
A useful companion for thinking about tool surfaces, discovery, and how agents should connect.
Read more →Field note
The security posture that should surround any agent-readable web surface you expose.
Read more →