Ahrefs Site Audit: Setup, Health Score, Patches and Limits

What Site Audit does, and what it costs you

Ahrefs' Site Audit crawls a website and checks it for 170+ technical and on-page SEO issues ^[1]. It is free for any site you can verify, which makes it one of the most generous free audit tools available. This guide walks through setting it up, reading it and acting on it, based on Ahrefs' own help center as of 12 October 2026. The short version is simple. Create a project, verify ownership, choose URL sources and crawl settings, then run the crawl. Read the Health Score, then work through Errors, Warnings and Notices in that order. Everything else in the tool supports that loop.

The currency of Site Audit is the crawl credit. Verified sites get 5,000 free crawl credits a month each, and paid plans add up to 5 million ^[2]. The paid tiers differ mainly in credits for other sites, the size of a single project and access to HTTP authentication:

What actually spends a credit

Only internal HTML pages that return a 200 status spend credits. Redirects, 4xx and 5xx URLs, images, CSS, JavaScript and external links are crawled for free ^[2]. That is generous. A site with many broken pages does not pay for them, so a store with 20,000 live product pages and 30,000 redirects and dead URLs spends 20,000 credits per full crawl, not 50,000. You can also cap the internal pages per crawl, which caps the credits any single crawl can spend. Sites larger than 5,000 pages need Starter or higher, and paid plans let you audit sites you do not own, such as a prospect's site before a sales call. Our Ahrefs pricing guide covers the plans in full.

Setting up the first crawl

Verify the site before anything else

Verification unlocks the free credits and faster crawling. You can verify through Google Search Console, an HTML file, a DNS TXT record or a meta tag, and WordPress sites have their own HTML tag instructions. The Search Console route has a bonus: it also keeps GSC data beyond Google's 16 months. Do it first. It costs nothing, and every later setting depends on it.

Choose where the crawler starts

URL sources are the seed pages the crawler visits first. By default, Site Audit starts from the website itself and the sitemaps listed in robots.txt, but you can add specific sitemaps instead. To crawl only a sitemap, select only Specific sitemaps and set the maximum depth from seed to 0.

Set a speed your server can take

By default, Site Audit crawls up to 30 URLs a minute, and on verified sites you can raise that to 30,000 a minute. The right speed depends on your server: budget shared hosting should stay low, while dedicated servers can go higher. Too fast, and real visitors may notice. Too slow, and big crawls take days. At the default 30 URLs a minute, a 100,000-page site needs more than two days of continuous crawling, which is why verification and a higher limit matter for large sites.

Ahrefs' own Site Audit page promises "Up to 170k URLs crawled each minute", while the help center sets the maximum for a verified site at 30,000. The two figures may measure different things, such as total crawler capacity versus one project's limit. For planning, use the help center's 30,000.

JavaScript, exclusions and the schedule

If your site is built with JavaScript, enable Execute JavaScript, or Site Audit may see only a few pages. JavaScript rendering has been available on every plan since 2020, and Site Audit can then search raw HTML, rendered HTML and extracted text. Expect slower crawls with it on.

Next, trim what you do not need. You can include or exclude URLs by pattern with regular expressions, and Remove URL parameters crawls example.com/help/ once instead of every ?query= variant. Noindex pages cannot be excluded before a crawl, because the crawler only learns about noindex by visiting the page, so exclude them from reports instead. Finally, scheduled crawls can run daily, weekly or monthly, at a chosen time and time zone, and may start any time within the chosen hour. Ahrefs notes that most users crawl weekly, on Fridays. Pick a quiet time for your server.

Mobile, staging and blocked crawls

Site Audit can crawl the mobile version of a site with a preconfigured mobile user agent, and it can crawl staging and password-protected sites once you add credentials. HTTP authentication for that is on Advanced and higher, so audit staging before a launch, not after, if your plan allows it. When the crawler cannot get in at all, check robots.txt for rules blocking AhrefsSiteAudit or AhrefsBot. Firewalls, CDNs like Cloudflare and security plugins can also block it, and Ahrefs publishes its IP ranges for whitelisting. A www mismatch can stop a crawl too.

Reading the Health Score without fooling yourself

Health Score is the share of crawled internal URLs without Error issues, as a percentage ^[3]. Only Errors lower it; Warnings and Notices do not. Ahrefs rates 0 to 30 as Weak, 31 to 70 as Fair, 71 to 90 as Good and 91 to 100 as Excellent. The math is plain: ten pages with two errors score 80. That makes the number easy to explain to a client, and it also tells you where the effort belongs. Errors are the most important and need fixing as soon as possible. Warnings are less critical but still need a fix. Notices are nudges, for later.

Do not game the score

Turning off an Error issue raises the Health Score, but the problem is still there. Ahrefs says so plainly: Site Audit simply stops counting it. Use the setting only for issues that genuinely do not apply to your site, such as a deliberate noindex section. Write down why you turned each one off. Six months later, a new team member will see a perfect score and assume the site is clean, when in fact the problem was simply hidden from the calculation.

Custom issues for your own rules

You can change the importance of any pre-set issue and create custom issues from Page Explorer filters. Ahrefs' own example is a custom issue for meta descriptions longer than 160 characters. You build the filter in Page Explorer, click Create an issue, then name it, choose a category and set its importance, and it can apply to all new projects. Use custom issues for your own style rules. They then sit in the same list as the built-in checks.

The reports that catch most real problems

Indexability, robots.txt and sitemaps

Site Audit treats a page as indexable when it is HTML, returns 200, has no canonical to another URL and has no noindex tag. Many reports filter on that flag. A page that fails any of the four will not be indexed as itself, and the most common surprise is a canonical tag left pointing at a staging or old URL after a redesign, which quietly removes the live page from search. Site Audit also reports whether robots.txt timed out or blocked URLs, and whether each URL was in a sitemap. Pages in a sitemap that are noindexed, redirected or broken each appear as issues. A clean sitemap lists only indexable pages that return 200.

Duplicates and orphan pages

Duplicates are pages with the same or similar titles, descriptions, H1s or content. They are bad only when hreflang or canonical tags do not show the relationship between them, and the Duplicates report charts unique pages, good duplicates, bad duplicates and empty pages. Start with the site-wide fix. If a whole site exists in both HTTP and HTTPS, one redirect from HTTP to HTTPS fixes every duplicate. For duplicate pages, delete, noindex or canonicalize them; for duplicate titles on different pages, write unique ones.

Orphan pages have no internal links pointing to them. Be careful here, though. Site Audit can report pages as orphans when it found them only in a sitemap, even if links exist elsewhere. Check before acting. A true orphan needs at least one link from a relevant page.

Hreflang, structured data and page speed

International sites can view hreflang issues as a network graph, and Site Audit flags missing return tags, links to non-canonical pages and language mismatches as separate issues. Fix them in templates, not page by page. Hreflang tags are almost always generated by the same template across every language version, so one template fix often clears thousands of URLs.

Structured data is validated against Schema.org and Google's rich result rules, with the two sets of errors shown as separate issues, and Page Explorer can filter by schema items, properties, values and issues. For a store, check Product markup first, since missing price or availability properties can remove the rich result that makes a listing stand out in Google. For speed, a free Google PageSpeed Insights API key adds Lighthouse and Core Web Vitals data, from the next crawl on.

Digging deeper into the crawl data

Page Explorer and Link Explorer let you filter crawled data on more than 250 data points per page. The URL details panel shows content fields, inlinks, outlinks, hreflang, structured data and duplicates for one URL, and you can search raw HTML, rendered HTML and text without setting up a custom search before the crawl.

Internal linking gets two tools of its own. Page Rating measures a URL's internal and external link strength on a logarithmic 1 to 100 scale, and each page splits its rating equally across the pages it links to. It is computed only after the crawl finishes, and it shows which pages your own internal linking favors. Better still, Site Audit suggests internal links automatically: which page to link from, which to link to, with the keyword and context. Ahrefs' free tools guide calls the Internal link opportunities report a hidden gem. It is one of the quickest ranking wins in the tool.

Two views help on larger sites. The Site Structure report shows subdomains and subfolders as a tree, with status codes, link depth, organic traffic and referring domains for each, so you can find sections that attract links but get little traffic, or sit too far from the home page. Clients understand the chart view quickly. Segments isolate one part of a site, such as a subfolder or subdomain, and you can save one and return to it in a few clicks. Review the blog and the store as separate segments. Their problems rarely overlap.

Tracking change over time

Site Audit compares any two crawls on health, issues, and new and removed URLs, and a chart shows the health of up to 10 consecutive crawls. The What's new tab lists issues that grew since the last crawl. That tab is the fastest weekly check. Five minutes.

Plan your records around one hard limit. All plans keep Site Audit data for 12 months only, and older crawls are deleted. If you need a longer record, export a PDF overview and a CSV of issues. Agencies that report year over year should export every quarter. Once a crawl ends, you can download every issue as a ZIP or CSV with fixing instructions included, and send it to developers as is. On credit plans, Site Audit does not use report credits at all. Only crawl credits apply.

Always-on audit and alerts

Always-on audit crawls continuously, prioritizing important pages, and emails alerts for critical changes. It shows the site's state today and lets you jump to any past day. It is on for Lite and higher and needs a boost on Free and Starter; Pro boosts raise its speed to 10 URLs a minute, Max to 30. Since January 2026, its alerts can be tuned per issue by severity and site size, and changed in bulk, while an activity feed explains each change with a short AI note. Start with alerts for Errors only. Add Warnings once the feed is quiet.

Fixing issues from inside Ahrefs

Patches fix issues from inside Site Audit, without a developer. Ahrefs describes them as "a quick and temporary fix", and they can be drafted, published, unpublished and traced in a history log. They need the Max Project Boost at US$200 a month per project ^[4]. What they can change depends on which page you read. The help center says the first release patches only titles and meta descriptions, while the Patches product page now lists canonicals, redirects, noindex and nofollow tags, alt text and internal links as well. Trust the product page. The help article is older.

The AI helpers

Ask AI writes titles and meta descriptions for flagged pages, and Batch AI does it for many pages at once, limited to 1,000 pages per batch and 50,000 pages per project a month. Since May 2026, a Fix with Agent A button sends issues to Ahrefs' AI agent, and Copy for AI exports them as Markdown. Site Audit can also flag pages with a high share of AI-written content. On Pro-boosted projects, bulk detection is capped at 1,000 pages per crawl, and Max removes the cap. Publishers checking freelance content find it useful. Treat a flag as a reason to read the page, not as proof.

Running Site Audit for many sites

Agencies get two features that matter at scale. Since early 2026, project settings can be updated in bulk: you select several projects, choose which settings to override and apply them at once. Agencies with dozens of client sites will feel the difference, so set a standard crawl template first. The API, meanwhile, includes Site Audit endpoints for projects, page content and Page Explorer data, and early 2026 added an endpoint to filter Site Audit projects. That lets an agency pull issue counts into its own dashboards instead of logging in to each project. Each request costs API units.

A weekly audit routine

Once the first crawl is clean enough to work from, a light routine keeps it that way. This one keeps a site healthy in an hour a week:

• Monday: read What's new and fix new Errors.
• Wednesday: check orphan pages and broken internal links.
• Friday: review Patches drafts and publish.
• Monthly: export the crawl for your records.

The bottom line

Site Audit is excellent and, for verified sites, free. Its Health Score is easy to read and easy to game, so use Errors as the to-do list and keep a note of every issue you switch off. Verify first, set a sane crawl speed, and export before 12 months pass. Add Always-on audit or Patches only where they save developer time. For the wider picture of what the free tier covers, read is Ahrefs free, and for the whole toolset, our Ahrefs review.