Skip to main content
Site Audit maps your website, checks it for problems that stop crawlers and AI answer engines from using it, and scores each page. Open it from Website → Site Audit in the sidebar. It needs an active or trialing plan. Viewers can see results but can’t run audits.

How an audit works

An audit has two steps. The first is free. The second uses credits.
1

Run site audit: find your pages and check the site

The first time you open Site Audit, click Run site audit. ChatFeatured reads your sitemap and follows internal links to list your pages, and runs the site-wide checks (robots.txt, sitemap, llms.txt, HTTPS). It changes nothing on your site and doesn’t score pages yet.After the first run, the button reads Find new pages. Use it when you’ve added pages or changed your robots.txt or sitemap.
2

Audit pages: score each page

Click Audit all N pages to open the Run page audits sheet. It shows the scope, the estimated time, and how many credits the run uses. Click the audit button in the sheet to start.To audit part of the site, pick a folder in the site tree first (the button becomes Audit /blog (N pages)), or select rows in the table and click Run N audits.
A single page usually finishes within a minute. To audit a page the crawl missed, click Add pages and paste the URLs.
The Run page audits sheet, showing the website, scope, estimated time, the routes included and the credits the run uses

Page audit credits

Every page audit uses one credit, including re-audits. Credits are shared across the workspace and reset with your billing cycle. If a run needs more credits than you have, the sheet warns “Some pages will be skipped” and audits as many as your balance covers.

Scores

Each audited page gets three scores from 0 to 100:
  • Technical: can crawlers reach, load and index the page (status, redirects, HTTPS, links, canonical and robots tags).
  • Content: how the page presents itself and reads (title, description, social tags, length).
  • AEO: can AI answer engines read and quote it (structured data, direct answers, citation signals, content that loads without JavaScript).
Every category starts at 100, and each open issue takes off points by severity: 25 for Critical, 12 for Warning, 5 for Info. Each issue shows a chip such as +12 Content with the points you’d gain by fixing it. The Site score is a weighted average of your audited pages’ scores: Technical 40%, Content 35%, AEO 25%. Site-level checks appear in Findings but don’t change any score. Scores of 90 and up are Good, 70 to 89 Fair, 50 to 69 Poor, and below 50 Critical.

What it checks

Site-level checks run on every site audit: robots.txt and sitemap reachable and valid, sitemap listed in robots.txt, a valid HTTPS certificate with HTTP redirecting to HTTPS, AI crawlers allowed, llms.txt present and valid, and unique page titles and meta descriptions across your audited pages. Page-level checks run when you audit a page. They cover the page’s status code, indexability (including robots.txt), canonical and redirects, broken internal links, mixed content, title and meta description, Open Graph tags, content length, JSON-LD structured data, whether the content loads without JavaScript, heading structure, answer format, citation authority, and whether the page can be quoted in AI answers.

AI crawlers in robots.txt

The AI Crawlers Allowed check sorts crawlers into two groups, because blocking each costs you something different.
  • Answer crawlers fetch pages to answer questions or build a search index: OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot, Googlebot and Bingbot. Blocking any of them takes your site out of that product’s answers, so it’s critical.
  • Training crawlers collect training data, such as GPTBot, ClaudeBot, Google-Extended, Applebot-Extended and CCBot. Blocking only these is reported as info, since it doesn’t affect where you show up in AI search. The exception: blocking Google-Extended also stops the Gemini app grounding answers in your site.

llms.txt

When an llms.txt check fails, click Generate LLMs.txt fix. ChatFeatured drafts an llms.txt for your site. Copy or download it, publish it at the root of your website, then click Re-run checks.

Explore your site

Site Audit opens on the Explore view. The site tree on the left narrows everything to a folder, and the header shows the scores and issue counts for that scope.
The Site Audit Explore view, with the site score, category scores and open issues in the header, the site tree on the left and the Pages table
  • Pages tab: every discovered page. Filter by status, issues, page type and (with Search Console connected) index status. The actions menu has Columns, Export pages CSV, Export issues CSV and Share view.
  • Findings tab: everything that’s open. Start with these pages suggests where to begin, weighted by traffic when Search Console is connected. Site-level checks and Page issues list the open issues, with Add all to send them to Tasks.
Click a page to open its details: its open issues with how to fix each, the searches that find it (with Search Console connected), and its title tag and meta description against the recommended lengths. Click Re-audit to check it again, which uses one credit. If the site is verified in Index Submissions, Request indexing submits that URL.
An audited page, with its Technical, Content and AEO scores and the open issues under each

Act on an issue

Each open issue has three actions:
  • Fix with AI opens the agent in a side panel with the issue and the page loaded.
  • Mark as fixed counts the issue as passed straight away. The next audit checks it again.
  • Assign creates a task for a teammate.
To send many issues at once, use Add all issues on a page or in Findings. When a re-audit finds an issue fixed, its task completes on its own.

Multiple sites

The first breadcrumb next to the Site Audit title is the site switcher. Its menu has Add site, Reorder sites and All websites. Subdomains such as blog.example.com are separate sites; www.example.com and example.com are the same site.

Troubleshooting

The domain you entered for a new site didn’t resolve. Check the spelling and enter just the domain, such as example.com.
Your site’s firewall or bot protection blocked our request, so the file may exist. Crawlers can hit the same wall. Allow automated requests to /robots.txt and your sitemap.
Click Find new pages to re-read your sitemap, or use Add pages to add the URL directly.
Last modified on October 22, 2026