Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.zued.ai/llms.txt

Use this file to discover all available pages before exploring further.

zued doesn’t just look at your URLs as a list of separate pages. It builds two layers of understanding on top of them — topics and entities — and uses those to decide which prompts go to which URL, which pages compete with each other, and where your content has a gap.

​
Topics

A topic (also called a content cluster) is a group of URLs that cover the same subject matter. If five of your pages talk about retirement planning from different angles, they belong to one topic, not five. Topics matter for three reasons:
  • Prompts are generated per topic, then assigned to a URL. A prompt about “best retirement product for a freelancer” is generated once for the topic, then routed to the single URL that covers it best. Without topic grouping, you’d get duplicate prompts on every page that mentions retirement.
  • Cross-URL coverage is measured per topic. When you have several URLs in the same topic, zued compares how well each one covers different angles of the topic, so you can see overlap and gaps. See Cross-URL Analysis for the actionable view.
  • Recommendations reuse topic context. When zued suggests a section to add to a page, it draws on what the rest of the topic already covers — so the recommendation fills a gap rather than duplicating a sibling page.
Topics stay stable across re-crawls and across languages. If you add a German URL that covers the same subject as an English one already in the project, both end up in the same topic — not in two parallel clusters.

​
Entities

An entity is a concrete thing zued has identified on your pages: a product you sell, a feature you describe, a place you mention, a concept you teach, a person you reference. zued reads each page and extracts the entities that show up, along with the relationships between them. Entities matter because they’re how zued knows what your content is actually about. Two pages can talk about “retirement” in completely different ways: one about products, one about regulation. The page-level word “retirement” doesn’t tell you that — the entities do. zued uses entities to:
  • Group URLs into topics. URLs that share entities and entity relationships end up in the same topic.
  • Match prompts to chunks. When a prompt mentions a concept, the chunk on a page that contains the same entity is a more reliable match than one that just shares vocabulary.
  • Spot naming inconsistencies. If a single product appears on your page under three different names (“savings account”, “Sparkonto”, “Sparbuch”), AI engines may treat them as three different things. zued surfaces these naming variants so you can decide whether to standardise.
  • Drive page recommendations. When a recommendation suggests adding a section, it grounds the example phrasing in entities you already use — so the suggestion sounds like your content, not a generic template.

​
Why this matters for your audit

Without topics and entities, an audit is just a flat list of pages with prompts attached at random. With them, the audit knows:
  • Which pages compete with each other for the same AI engine question
  • Which pages should be linked together because they cover related entities
  • Which prompts are on the wrong URL because the entity they’re about lives elsewhere on your site
  • Which entities are mentioned without being explained, and which are explained well enough that AI engines can cite them
You see the output in three places: the Cross-URL Analysis panel for sibling-page coverage, the page-level recommendations on each URL, and the match confidence label on every prompt (see Prompts).

​
A note on stability

Topics and entities are recomputed when zued has reason to believe your content has changed meaningfully — for example after a re-crawl with a large content delta, or when you add or remove URLs from a project. Day-to-day, they stay stable. This is intentional: the comparison week to week needs the same scaffolding, otherwise trends become noise.