---
title: "Product facts"
description: "What Blogged knows about your own product, where it came from, and how to correct something that is wrong."
updated: "2026-08-18"
verifiedAgainst:
  - packages/shared/src/knowledge.ts
  - packages/shared/src/site-analysis.ts
---

Your product is an entity of kind `own_product`, and the facts attached to it are
the ones that appear most often in your writing.

## Where they come from

| Origin | Meaning |
| --- | --- |
| `onboarding` | The evidence you gave at setup |
| `official_crawl` | A read of your own public pages |
| `user` | Something you stated yourself |
| `exa` | A market search performed on your behalf |

Facts extracted from a page record how they were extracted, from structured data
in the page, from the page's own markup, or from a model reading it. That
detail is what makes a wrong fact diagnosable rather than mysterious.

## Facets

Facts are filed under six facets: `pricing`, `features`, `integrations`,
`limits`, `positioning` and `guidance`. The facet is what a drafting run asks
for, so a comparison article pulls pricing and limits while a how to pulls
guidance.

## Correcting a fact

State the correct version yourself. A fact you confirm carries a stronger status
than one extracted automatically, and it is preferred over a crawled fact that
disagrees with it.

<Callout variant="tip">
  If your public pages are the thing that is out of date, fix them and refresh
  the source. Correcting the fact by hand while the page still says otherwise
  leaves a conflict that will keep coming back.
</Callout>

## Pages worth having

Blogged classifies the pages it reads: homepage, pricing, features,
integrations, documentation, security, changelog, about, comparison, blog. A
product with a real pricing page and real documentation gives the engine
materially better facts than one with a landing page and a signup form.