truthbased.
Start free

Public truth surfaces

Your Truth page, your llms.txt and your entity map. What each one holds, who fetches it, and how to point your own site at them.

5 min read · Updated 2026-10-05

The three addresses

Publishing your Truth page creates three public addresses. They serve the same signed facts in three shapes.

AddressForWhat it holds
/t/your-brandPeople, and crawlers that render pagesThe Truth page: every signed fact with its source and date, grouped by topic, with search.
/t/your-brand/llms.txtAI crawlersA plain-text digest: your brand article, then one line per fact with its verification date and trust score, then how to reach the MCP endpoint.
/t/your-brand/entitymap.jsonMachinesAn EntityMap v1.0 file: your brand as an entity, its topics, and for each a handful of quoted chunks with source, date and relevance.

All three return a real 404 until the brand is published, and they refresh within five minutes of a change. The Truth page is indexable and is listed in Truthbased's own sitemap.

  1. Open the brand, then Connections, then the Public truth surfaces tile. The drawer lists the three addresses as links.

    The Truth surfaces drawer. ① The human-readable Truth page. ② The entity map for machines. ③ The plain-text digest for LLM crawlers.
    The Truth surfaces drawer. ① The human-readable Truth page. ② The entity map for machines. ③ The plain-text digest for LLM crawlers.

What llms.txt looks like

Your llms.txt as a crawler sees it: the article, one line per fact, then the MCP endpoint.
Your llms.txt as a crawler sees it: the article, one line per fact, then the MCP endpoint.

Facts are sorted by trust. A fact that is no longer current is tagged historical or deprecated rather than removed, so a crawler can tell the difference between "was true" and "is true".

What the Truth page tells agents

Near the top of the public page, a dark panel called "For AI agents" links to llms.txt and the entity map and names the MCP tools. A crawler that lands on the page finds the machine files from there, and the page's own head declares them as alternates too.

The public Truth page. ① The For AI agents panel.
The public Truth page. ① The For AI agents panel.

Point your own site at them

AI crawlers look for machine-readable facts on your domain. Three snippets, all plain HTML with no script, tell them where the record is. They are under Connections, then the Static snippets tile.

The Static snippets drawer. ① Head tags. ② The llms.txt line. ③ The footer badge.
The Static snippets drawer. ① Head tags. ② The llms.txt line. ③ The footer badge.

  1. Head tags. Two <link rel="alternate"> lines for your <head>, pointing at your llms.txt and entity map. Invisible to visitors, read by crawlers. If you installed the Truth Tag with discovery links on, you already have these.

  2. The llms.txt line. One line to add to your own site's /llms.txt. Create the file at your site root if you do not have one; AI crawlers check for it.

    - [Verified brand facts](https://truthbased.ai/t/your-brand/llms.txt): human-governed, source-cited facts about Your Brand, maintained with Truthbased

  3. The footer badge. Optional and visible: a small "Verified facts" seal that links to your Truth page. The quickest way for a crawler to find the page from your site.

Press Re-check my site on Connections afterwards. The tile reads Connected when your homepage carries the head link or your JSON-LD. It cannot see inside your llms.txt, so that line stays on your word.

If something is off

  • A surface returns 404. The brand is not published, or the address has a typo. Check Settings, Truth Center.
  • The page shows "sample" and ten facts. The plan does not include the full feeds. The Truth page itself is complete on every paid plan.
  • A fact you corrected still shows. Wait five minutes for the cache, then reload.

What to do next