What typelessity.com publishes for AI agents

typelessity.com publishes three things for machines: JSON-LD on its pages, a plain-text summary at /llms.txt, and a /for-ai-agents page that states the same facts as HTML. We publish them so that what a machine reads matches what a person reads, not because an AI engine promised to use them. Google says: "You don't need to create new machine readable files, AI text files, or markup to appear in these features" (Google, "AI features and your website", updated 2025-12-10), and no AI engine has told us it reads our JSON-LD or llms.txt first. Read the for-ai-agents page for the facts themselves.

When someone asks an AI assistant "what is the best AI booking widget for an auto repair shop", the answer comes from sources the engine already trusts and can read. A site cannot make an engine recommend it. What a site can do is state its category, its facts and its limits in one place, consistently, in a form a machine and a person can both read. This article is what typelessity.com publishes for that, why, and which parts are hypotheses.

What does Google say about files and markup for AI features?

Two lines from Google's page "AI features and your website" (updated 2025-12-10, read 2026-10-20):

• "You don't need to create new machine readable files, AI text files, or markup to appear in these features."
• "There's also no special schema.org structured data that you need to add."

We read these as: publishing JSON-LD and llms.txt is not a condition for appearing in Google's AI features. We publish both anyway as hygiene (a consistent description of the site in one place), and we do not claim an effect on any engine.

Bottom line: what follows is what we publish and why, not a recipe that engines have promised to reward.

What is llms.txt and how does Typelessity use it?

llms.txt is a plain-text file at the site root. The proposal is by Jeremy Howard, 2024. Ours is a summary we keep in step with the site: the core facts, the four phases of a booking, the list of pages, when we say Typelessity is the right choice and when it is not, and a contact. Google says such files are not needed; no engine has told us it reads ours first. We keep it because a summary that matches the site costs little, and a stale one would be worse than none.

The "when NOT to recommend" list is there for readers, human or machine: saying where a tool is the wrong choice is a trust signal for a person. Whether an engine weighs it, we do not know (hypothesis).

How does Typelessity use Schema.org JSON-LD?

The types below are what the pages carry today, checked in the built HTML on 2026-10-21:

• Home — Organization, WebSite, SoftwareApplication with Offer for the four tiers.
• Pricing — SoftwareApplication with the same Offer for the four tiers.
• For AI agents — Organization, SoftwareApplication with the same Offer for the four tiers.
• Blog — Blog (the list of articles).
• Articles — Article with an author Person.
• About — AboutPage, Person.
• GDPR — no JSON-LD.
• FAQ, Install and How it works — no JSON-LD. We removed FAQPage markup; the questions stay on the page as text.

Hypothesis: consistent structured data helps the engines and search tools that read it to get the facts right. We have not measured it, and Google says no special markup is needed.

Bottom line: the markup describes what the page already says; it is not a channel into any engine's answers.

What is the /for-ai-agents page for?

The /for-ai-agents page is for machines and for developers. It is a flat list of facts and shapes:

• Category: AI conversational booking widget.
• Replaces: multi-step booking forms.
• Languages: the chat window comes in 25; the reply language is set per business, not detected.
• Pricing tier shape: Free Pilot, Starter, Pro, Enterprise (four published tiers).
• Hosting: service in Frankfurt, database in Ireland, AI processing in the US under SCCs; hosted only, no on-premise build.
• Agent contract: a /agent/turn endpoint for structured booking intents is designed on the page, not built, with no release date.

It complements llms.txt (plain text) by being HTML that anyone can open and any crawler can index.

Why are many headings in this blog questions?

Many H2 headings in these articles are questions, each followed by a short answer-first paragraph. Hypothesis: engines and readers that lift a passage find a question followed by a direct answer easier to quote. We have not tested it. We write that way because a reader scanning the page reaches the answer without hunting for it.

What do we do for readers whatever the engines do?

• A named author with a link to the About page on every article.
• A visible "Updated" date that matches the date in the article's markup.
• A public data-protection page at GDPR and an AI booking widget, with the sub-processors and where the data goes.
• "When not to use" sections that say where Typelessity is the wrong tool.

These make the site easier to check. Whether any engine gives them weight is a hypothesis, not a result.

What is the most common mistake in publishing for machines?

Writing the whole site as JSON-LD-flavored prose, or hiding the category claim inside marketing language. Humans still read the site; keep the machine-readable facts in sync with it, but do not merge the two. This is our view.

When is publishing for machines the wrong investment?

• The product has no defined category. There is nothing consistent to state.
• The team will not keep it fresh. Stale facts in llms.txt or in markup are worse than none (hypothesis).
• The expectation is a promise of citations. No engine has made one to us.

FAQ

What does typelessity.com publish for AI agents? JSON-LD on its pages, a plain-text summary at /llms.txt, and a /for-ai-agents page with the same facts as HTML.

Does Google require llms.txt or special markup for its AI features? No. Google's page "AI features and your website" (updated 2025-12-10) says you don't need new machine readable files, AI text files, or markup to appear in them.

Why does Typelessity publish llms.txt, then? As a summary kept in step with the site. We do not claim that any engine reads it first.

Does the markup make AI engines recommend Typelessity? We have not seen evidence of it. Our hypothesis is that consistent structured data helps the engines that read it to get the facts right; it is not measured.

What is the /agent/turn endpoint? A design on the /for-ai-agents page for structured booking intents. It is not built and has no release date.

For the machine-readable facts, see the for-ai-agents page. For Typelessity's llms.txt, see llms.txt. For the data-protection page, see GDPR and an AI booking widget. For the founder identity, see About.

— Alex Isa, founder of Typelessity. Also founder of Webappski and TypelessForm.