Page audits

The Page audits section scores any public page the way answer engines experience it: fetched with a real AI crawler identity, checked across eight categories, and stored so every re-audit shows what changed. The scoring is deterministic. The same page always produces the same result, and no category ever depends on a model's opinion.

Running audits

Audit a single URL, or a whole site as a domain audit: Citlyze reads the sitemap, takes the top pages up to your chosen cap, and audits each one in the background. Domain audits are saved as sets; switch a set to weekly and it re-audits automatically with a score diff against the previous run.

Pick the crawler identity the audit runs as. GPTBot is the default, and the crawlability category evaluates robots.txt for exactly that identity, so a page can score differently for GPTBot than for ClaudeBot when your robots rules treat them differently.

The eight categories

• AI-bot crawlability: robots.txt for the selected bot, retrieval-agent access, indexability, and response time. llms.txt presence is reported but never scored.
• Renderability: how much of the rendered page already exists in the static HTML. Most AI crawlers do not run JavaScript, so content that only appears after rendering is invisible to them.
• Structure and chunkability: one descriptive H1, clean heading hierarchy, question-shaped subheadings, an answer-first lead, tables and lists, and whether each section stays on its heading's topic.
• Citability: statistics density, quotable blocks, machine-readable dates, and vocabulary uniqueness. The signals that make a page worth citing rather than merely readable.
• Structured data: JSON-LD validity, whether the schema types match the page type, and metadata consistency.
• Authority: authorship signals, internal linking, and external source grounding.
• Fluff and subjectivity: marketing-speak density measured against a per-language lexicon, in all supported content languages.
• Performance: lab performance data for the page. Shown when performance data is available for your workspace.

Scoring is page-type aware: a product page weighs structured data more heavily, an article weighs citability, a support page weighs structure. The detected page type is shown on every audit. When a category cannot run for a given audit, it is excluded from the weighted total and labeled, never silently scored as zero.

Credits

Each audited page uses one credit from your plan's monthly pool; legacy GEO audits draw from the same pool. Failed audits are never charged. Agency plans add 200 credits per active client workspace, and the audit credit pack add-on extends any plan. When a scheduled set exhausts the pool it pauses with a visible notice and resumes next month.

Reports and history

Every audit is stored for twelve months. The detail view shows factor-level findings with evidence and a fix hint for anything below pass, plus the change against the previous audit of the same URL. The report button produces a printable report; on white-label plans it carries your own logo, accent color, and company name.