# Docs Audit (https://app.infrasity.com/docs/geo-dashboard/aeo-audit/docs-audit)





Docs Audit ("Documentation quality · AI-ready") is the AEO Audit tab built specifically for technical documentation, API guides, and code examples: the kind of content coding agents and developer-focused AI tools rely on most, and which has its own set of requirements beyond general marketing/content pages.

<Screenshot src="/images/geo-aeo-docs-audit.png" alt="Docs Audit tab showing a documentation URL input, a discovery method badge, an overall score with letter grade, Passed/Failed/Warnings/Manual counts, and a per-section results list" caption="Docs Audit: discovers your documentation pages automatically, then scores each section of the check." />

## What it's for [#what-its-for]

General-purpose SEO/AEO checks don't fully capture what makes documentation usable by AI models: whether code blocks are clean and properly labeled, whether API references are structured so an agent can parse parameters correctly, and whether explanations are self-contained enough to be extracted and cited without missing context. Docs Audit evaluates a documentation URL against exactly that criteria, the same underlying concern the [Dev Intelligence](/dev-intelligence/overview) module tracks from the traffic/citation side.

## Running an audit [#running-an-audit]

* **URL field + Audit Docs**: point it at your documentation (a live docs site or even a shared doc link).
* **Download PDF**: export the full result as a shareable report.

Docs Audit doesn't require a single-page URL: it discovers your documentation's other pages on its own, via `llms.txt`, your sitemap, or heuristic path probing if neither is available. A badge above the results shows which discovery method was used and how many pages it found.

## Score summary [#score-summary]

* **Score / Grade**: an overall score out of 100, mapped to a letter grade.
* **Passed / Failed / Warnings / Manual**: how the checks broke down. **Manual** checks are ones the tool can't fully verify automatically and flags for you to confirm by hand.

## Section results [#section-results]

Results are grouped into labeled sections. Each section shows its own fail/warn/pass counts, a progress bar, and a percentage score. Use the **All / Fail / Warn / Pass / Manual** filter to jump straight to the sections that need attention, and expand a section (▼) to see its individual checks.

* **AI & LLM Discoverability**: whether crawlers and agents can find your docs at all, `llms.txt`, sitemap coverage, and similar signals. *Why it matters*: none of the other sections matter if an agent can't locate the page in the first place.
* **Structure & Navigation**: headings, sidebar/page-tree structure, and consistent organization. *Why it matters*: agents often need to jump between related pages (a concept page and its API reference, for example); a predictable structure makes that traversal reliable instead of guesswork.
* **Content Completeness**: whether a page covers a topic end to end rather than assuming missing context. *Why it matters*: an agent extracting one page usually can't see the rest of your site; a page that depends on unstated prerequisites will get misapplied.
* **Content Quality**: clarity and precision of explanations. *Why it matters*: ambiguous wording is more costly in docs than in general content, since an agent may execute code based on a literal misreading.
* **Technical SEO & Crawlability**: baseline crawlability fundamentals applied to your docs specifically. *Why it matters*: documentation sites often live on separate subdomains or platforms from the main site, which can accidentally miss the crawlability setup the rest of the site has.
* **Internal Linking & Flow**: whether related pages link to each other in a way that reflects how the docs are actually used. *Why it matters*: good internal linking helps an agent (or a crawler building citation context) find the specific page a general question should route to.
* **Versioning & Maintenance**: whether the docs indicate what version they apply to and how current they are. *Why it matters*: an agent generating code from stale, unversioned docs can confidently produce integration code that no longer matches your current API.
* **Agent Access & Markdown**: whether pages are available in a clean, agent-consumable format (for example, raw Markdown) rather than requiring JavaScript rendering. *Why it matters*: agents work far more reliably from plain text than from a rendered HTML page they have to reverse-engineer, the same reason a `/llms.txt` file exists at all.
* **Agent Tooling & Stability**: whether code examples, install commands, and API references are formatted consistently enough for an agent to extract and run reliably. *Why it matters*: inconsistent formatting is a common reason an otherwise-correct agent produces broken setup steps.

<Callout type="gotcha" title="A low overall score doesn't mean every section is weak">
  Section scores in a real audit tend to vary widely: some sections can score well while others drag the overall grade down. Read the per-section breakdown rather than treating the overall score as evenly distributed across all of them.
</Callout>

<RelatedPages
  pages="[
{ title: &#x22;Domain-Wide Audit&#x22;, description: &#x22;Crawl your entire site and score every page.&#x22;, href: &#x22;/geo-dashboard/aeo-audit/domain-wide-audit&#x22; },
{ title: &#x22;Single Page Audit&#x22;, description: &#x22;Deep-dive diagnostic on one high-value URL.&#x22;, href: &#x22;/geo-dashboard/aeo-audit/single-page-audit&#x22; },
{ title: &#x22;AEO Audit&#x22;, description: &#x22;Analyze your content's answer engine optimization performance.&#x22;, href: &#x22;/geo-dashboard/aeo-audit&#x22; },
]"
/>
