readable, verifiable, citable
An agent-ready website
People ask ChatGPT before they search. If an agent cannot read your facts and verify them, your name never enters the answer. These pages explain each technical surface — and every one of them runs on this site.
The ten points
- 01
Crawlable by AI agents
robots.txt names the AI user agents and lets them in, instead of blocking them by accident.
- 02
A machine index
A short curated /llms.txt plus a full inventory, so a model can find the right page without guessing.
- 03
Markdown mirrors
Every page also exists as clean markdown, so an agent reads all of it instead of truncating the HTML.
- 04
One connected data graph
JSON-LD for the business, its services, articles, dates and FAQs, sharing ids so it resolves as one entity.
- 05
Facts as text, not pictures
Prices, hours, licences, menus and credentials written as text — agents do not read your poster.
- 06
Native Arabic with hreflang
Arabic pages written natively, not translated, and tagged so an Arabic question gets the Arabic page.
- 07
Agent-callable tools (MCP)
A read-only MCP server an agent can call for services, availability and local calculations.
- 08
In-page tools (WebMCP)
Booking and contact forms registered as tools a browser agent can fill, handing off to WhatsApp.
- 09
Honest freshness signals
Real modified dates and sitemap lastmod only where content actually changed, plus IndexNow pings.
- 10
Measured, not assumed
Referrals from ChatGPT, Perplexity and Gemini tracked in analytics, and every WhatsApp click counted.
Frequently asked questions
What makes a website agent-ready?+
The facts a customer asks about exist as plain text, the AI crawlers are allowed, there is a machine index and a markdown mirror of every page, the structured data resolves to one business, the Arabic is native with correct hreflang, and where it earns its place there are tools an agent can call.
Does this make ChatGPT recommend my business?+
No one can buy a place in the answer. This work makes your business readable, verifiable and citable, which is the precondition for being named. Being recommended is still earned through consistent facts, reviews and mentions on sites you do not own.
Do the AI engines promise to read llms.txt?+
No, and anyone who says otherwise is guessing. llms.txt is a published convention that a growing number of tools read. Markdown mirrors and clean structured data help every agent regardless, which is why we build those first.
Can I see it working somewhere?+
On this site. Append .md to any page and you get the markdown twin, /llms.txt is the curated index, and our MCP server answers agent calls. The pages below explain each surface and what it costs to add.
the next step is small
Want us to audit your site?
- Send your URL, what you sell, and who buys it.
- We check the ten points and tell you what is in place and what is missing.
- You get the fixes ranked, then decide who does them.