---
title: "Public Site"
id: "1855"
type: "page"
slug: "public-site"
published_at: "2026-09-21T00:23:49+00:00"
modified_at: "2026-10-22T16:38:23+00:00"
url: "https://xedant.com/agents/portal/docs/public-site/"
markdown_url: "https://xedant.com/agents/portal/docs/public-site.md"
excerpt: "Part of the portal can be opened to the outside — to ordinary visitors, without…"
---

# Public Site

[https://xedant.com/agents/portal/docs/public-site/.md](https://xedant.com/agents/portal/docs/public-site/.md)

Part of the portal can be opened to the outside — to ordinary visitors, without signing in. It is a separate site with its own look, built entirely from the portal pages with “public” access. The result is convenient: the same agent maintains the content, and visitors see a tidy site.

## What goes outside

Only pages with “public” access are visible outside. Everything else — drafts and staff pages — does not open from the outside and is not shown. More about access levels is in the [Page Access](/agents/portal/docs/access)
 section.

## Addresses

The public site lives at its own address: `/p/` is the site’s home page, `/p/news` a specific page. No sign-in is ever requested on either: the visitor simply opens a link from an email or a search result. A page’s portal address and its address on the site differ only by the leading `/p/`.

## The site menu

The menu assembles itself from the published pages. The order of the items and their labels come from the page settings, so there is nothing to rearrange by hand. The menu depth — one level or two with a dropdown — is set once in the home page’s settings. A page can also be hidden from the menu while staying reachable by direct link: handy for special offers or utility pages.

## For search engines

Every public page has a browser tab title and a description for search engines. Both come from the page itself, so the agent sets them together with the content. The title combines the page name and the site name, and the description comes from the page description. When a page has no description, the search shows the name only.

## What the public site does not have

The public site is deliberately simpler than the inner portal. It has no page tree, no agent chat, no recommendation card and no request-to-agent buttons — a visitor does not need working tools. The agent’s service messages are not shown either. The visitor sees only finished pages, the menu and buttons that lead to pages or run the configured actions.

## A hidden page and a non-existent page look the same

When a visitor opens the address of a closed page, they get the same tidy “Page not found” page as for a typo in the address. The answer gives no way to tell whether the page really exists. This protects drafts from curious eyes: a section being prepared does not give itself away, neither by the page’s look nor by the response text.

## Speed and data freshness

Images on the public site are remembered by the browser for an hour — a repeat view is noticeably faster, and the image is not downloaded from the server again needlessly. Live block data updates itself when the visitor returns to the site’s tab: while it was in the background, there is no reason to show stale numbers.

## Forms and protection from frequent submissions

Forms and navigation on the public site work the same way as inside the portal: a visitor can subscribe to news, leave a request or move to another page. To keep robots from abusing a form, no more than ten submissions a minute are accepted from one address — the rest are politely declined, and the page shows a readable message. Details are in the [Forms](/agents/portal/docs/forms)
 section.

## Live numbers and files outside

A public page can show your data as well: a visitor sees the same summaries as the staff do inside the portal. The data arrives by a query description that lives in the page itself — the SQL is defined by the document, never by the reader (see the [Data Sources & Visualisations](/agents/portal/docs/sources)
 section). A reader gets block data through their own address, and a draft or closed page answers “no such page” and computes nothing: publishing a page also publishes the results of its queries.

Pictures and files from a page’s folder work outside just as they do inside: a picture shows right in the text, and a file downloads by its link without taking the visitor to another page.

Next: where live numbers come from — in the [Data Sources & Visualisations](/agents/portal/docs/sources)
 section; how to compute data with a script — in [Dynamic Python Blocks](/agents/portal/docs/python)
; how to open or close a page — in [Page Access](/agents/portal/docs/access)
; how to connect the agent to the chat — in [Chat with the AI Agent](/agents/portal/docs/agent)
.

[← Back to the documentation index](/agents/portal/docs)