How it works

Minutes of setup. None of it is scripting.

You never write a canned answer or draw a decision tree. You point Sitora at the site you already maintain.

How it works

From URL to working agent in minutes

Three steps, none of which involve writing a single answer by hand.

step 1

Paste your URL

That is the whole setup. No uploads, no retyping your FAQ into a dashboard, no training sessions.

sitora reads pages, docs, FAQs, pricing
step 2

Sitora reads your site and grounds itself

Pages become passages, passages become the only material your agent is allowed to speak from. If it is not on your site, your agent does not say it.

grounded passages with page-level sources
step 3

Embed with one script tag

Drop one line into any site builder or template. Your agent answers customers where they already are: on your site.

<script src="https://getsitora.com/embed.js" ...>
What gets read

Your public pages, within honest caps

Sitora reads the pages a customer could read: home, services, pricing, FAQs, contact, and the pages they link to. Content becomes retrievable passages tied to their source page.

  • Up to 6 pages per read on the demo and free tier, prioritized by usefulness (pricing, services, FAQ, contact first).
  • Text only: no scripts run, no logins attempted, no forms submitted.
  • Re-read any time from the studio when your site changes.
  • Everything read is visible to you in the coverage view, and deletable.
Embedding

One tag, any builder

From the studio, copy your embed line and paste it before the closing body tag, in a footer widget, or in your tag manager.

<script
  src="https://getsitora.com/embed.js"
  data-agent="YOUR_AGENT_ID"
  async>
</script>

Works on WordPress, Squarespace, Wix, Webflow, Shopify, Framer, and hand-written HTML. The widget lazy-loads, keeps your page fast, and opens into the same grounded agent you tested in the studio.

Go deeper

Want the internals?

Retrieval, refusal thresholds, the answer contract, and what happens when the model is unsure.