What it reads
Four questions, asked as a machine would.
Each is a plain test of your public site: a handful of requests and a read of what comes back. Nothing is guessed, and every shortfall comes with the text to correct it.
- 01
Access
Whether ChatGPT, Claude, Perplexity, Google and Bing can reach your pages, or are turned away by robots.txt or a firewall before they read a word.
- 02
Content
Whether your pages carry their text in the HTML itself, with a clear title, description and outline, and are open to being quoted.
- 03
Structured data
Whether your name, your profiles and your offering are stated in schema.org markup, so machines need not guess who you are.
- 04
Agent discovery
Whether an AI agent can find what you offer: llms.txt, markdown on request, and the well-known files agents look for.
What it cannot tell you
Readable is not the same as recommended.
This check shows whether your site is open to AI systems. It cannot show what they currently say about you: whether they name you, recommend you, cite you, or describe you accurately.
That is a different question, answered by asking the systems themselves, across the questions your clients actually put to them.
The check requests your homepage, robots.txt, sitemap and a few well-known files, once each, identifying itself as JunctraReadinessBot. It does not log in, submit forms or read anything private.
Now ask the systems themselves.
Junctra puts the questions your clients ask to ChatGPT, Claude, Perplexity and Gemini, and reports exactly how each one describes you.