Technical SEO Page Audit

Audits public web pages for technical SEO and returns one JSON report per page: 44 fixed checks, each with a status, the evidence it was decided on, and a fixed recommendation when it warns or fails. The same HTML and response headers always give the same checksSha256.

It reads what a crawler reads: the server HTML, response headers, robots.txt and, if you ask, the status of the page's links. It does not run JavaScript and uses no AI model.

Input

Field	Type	Default
urls	array of http(s) URLs		Pages to audit, up to 100
sitemapUrl	URL		Optional sitemap or sitemap index; same-host pages are added in sitemap order
maxPages	1-100	10	Pages taken from the sitemap
checkLinks	boolean	false	HEAD-request the page's links and report 4xx/5xx
maxLinksToCheck	1-50	20	Links checked per page
timeoutSecsPerUnit	10-300	60	Time allowed per page

Give urls, sitemapUrl, or both. Example:

{
  "urls": [
    "https://example.com/"
  ]
}

Output

One dataset item per audited page:

• score.overall (0-100) and score.categories, computed from check weights: pass counts fully, warn half, fail zero
• counts of pass / warn / fail / info / not_applicable / not_checked
• checks[]: id, category, title, weight, status, evidence, recommendation (null unless the check warns or fails), measured (true for the two time-dependent checks)
• page: title, meta description, canonical, robots directives, H1s, word count, link and image counts, JSON-LD types, Open Graph, hreflang
• http: status, redirect chain, response time, HTML bytes, transfer bytes, content encoding
• completeness: whether robots.txt and the link check were fully read
• htmlSha256 (the HTML that was judged) and checksSha256 (the checks, excluding the two time-dependent ones)

A page that cannot be audited (DNS failure, timeout, 4xx/5xx, non-HTML, blocked address) produces an item with ok: false and an error.code instead. The key-value store record OUTPUT lists every page with whether it was charged and why.

Real output for the example input above, shortened to 3 of the 44 checks and 6 of the page fields:

{
  "url": "https://example.com/",
  "finalUrl": "https://example.com/",
  "ok": true,
  "score": {
    "categories": {
      "content": 84,
      "indexability": 93,
      "links": 75,
      "performance": 100,
      "security": 92,
      "structured_data": 50
    },
    "overall": 86
  },
  "counts": {
    "fail": 0,
    "info": 5,
    "not_applicable": 10,
    "not_checked": 1,
    "pass": 20,
    "warn": 8
  },
  "completeness": {
    "linkCheck": "skipped",
    "robotsTxt": "absent"
  },
  "htmlSha256": "25ddf2c883e0d1958ea971d279a7e4f0fd446724ee3db7db19dadabd4a62e484",
  "checksSha256": "80404755eac37451a50a6046a764bdb094592c32ead1074bc4bf41a6a4680935",
  "fetchedAt": "2026-10-08T02:51:10Z",
  "checks": [
    {
      "category": "indexability",
      "evidence": {
        "canonical": []
      },
      "id": "IDX004",
      "measured": false,
      "recommendation": "Add one <link rel=\"canonical\"> with an absolute URL in <head>; remove duplicates.",
      "status": "warn",
      "title": "Exactly one absolute canonical URL",
      "weight": 2
    },
    {
      "category": "indexability",
      "evidence": {
        "http_status": 404,
        "note": "no robots.txt: everything is allowed",
        "robots_url": "https://example.com/robots.txt"
      },
      "id": "IDX006",
      "measured": false,
      "recommendation": null,
      "status": "pass",
      "title": "robots.txt allows crawling this URL",
      "weight": 5
    },
    {
      "category": "content",
      "evidence": {
        "h1": [],
        "h1_count": 0
      },
      "id": "CNT005",
      "measured": false,
      "recommendation": "Use exactly one <h1> that states the page topic.",
      "status": "warn",
      "title": "Exactly one H1",
      "weight": 2
    }
  ],
  "page": {
    "title": "Example Domain",
    "metaDescription": null,
    "canonical": null,
    "lang": "en",
    "h1": [],
    "wordCount": 25
  },
  "http": {
    "status": 200,
    "redirects": [],
    "responseMs": 50,
    "htmlBytes": 577,
    "transferBytes": 400,
    "contentEncoding": "gzip"
  }
}

Pricing

Pay per event:

• page-audited: $0.02 per page that produced a report. A page is charged only after its report is validated against the output schema and saved to the dataset.
• Apify's run-start event, apify-actor-start: $0.00005 per run (one event: this Actor's memory is capped at 1 GB).
• Platform usage (compute, storage, data transfer) is not charged to you on top of these events.

Never charged: pages that fail to load (DNS, timeout, connection reset, 4xx/5xx, rate limiting), non-HTML responses, invalid or private-network URLs, and any page the run's maximum charge cannot cover (the run stops before fetching it). A report whose robots.txt or link check could not be completed is still a report and is charged; completeness says which part is missing.

Checks

Id	Category	Rule	Weight
IDX001	indexability	Final response is HTTP 200	5
IDX002	indexability	At most one redirect before the page	1
IDX003	indexability	Page is not marked noindex	5
IDX004	indexability	Exactly one absolute canonical URL	2
IDX005	indexability	Canonical points to this page	1
IDX006	indexability	robots.txt allows crawling this URL	5
IDX007	indexability	robots.txt declares a sitemap	1
IDX008	indexability	Response is HTML	3
CNT001	content	Page has a title	4
CNT002	content	Title length 10-60 characters	2
CNT003	content	Page has a meta description	2
CNT004	content	Meta description length 50-160 characters	1
CNT005	content	Exactly one H1	2
CNT006	content	Heading levels are not skipped	1
CNT007	content	At least 200 words of visible text	1
CNT008	content	declares lang	1
CNT009	content	Mobile viewport meta tag	2
CNT010	content	Character encoding declared	1
LNK001	links	Page links to other pages on the site	2
LNK002	links	Every link has an accessible name	1
LNK003	links	No broken links among those checked	3
LNK004	links	No placeholder links (# or javascript:)	1
IMG001	media	Every image has an alt attribute	2
IMG002	media	Images declare width and height	1
SD001	structured data	JSON-LD structured data present	1
SD002	structured data	All JSON-LD blocks parse	3
SD003	structured data	JSON-LD nodes have schema.org @context and @type	2
SD004	structured data	Open Graph title, description and image	1
SD005	structured data	Twitter card declared	1
INT001	international	hreflang values and URLs are valid	2
INT002	international	hreflang set includes this page	2
INT003	international	hreflang set has x-default	1
SEC001	security	Final URL uses HTTPS	5
SEC002	security	HTTP redirects to HTTPS	1
SEC003	security	Strict-Transport-Security header	2
SEC004	security	No mixed content	3
SEC005	security	X-Content-Type-Options: nosniff	1
SEC006	security	Content-Security-Policy header	1
SEC007	security	TLS certificate valid for 21+ more days	3
PRF001	performance	HTML document under 500 KB	2
PRF002	performance	HTML is compressed in transit	2
PRF003	performance	At most 2 render-blocking scripts in	2
PRF004	performance	Server answered within 800 ms	1
PRF005	performance	HTML response has Cache-Control	1

PRF004 (response time) and SEC007 (certificate days left) depend on when the page is fetched and are excluded from checksSha256.

What it does not do

• It does not crawl a site. It audits the URLs you give and, if you give a sitemap, up to maxPages same-host pages from it.
• It does not render JavaScript, measure Core Web Vitals (LCP, CLS, INP) or take screenshots.
• It does not log in, send cookies or fill forms; pages behind authentication are audited as an anonymous visitor sees them.
• It does not report rankings, keywords, backlinks or traffic.
• It does not change your site: every request is read-only (GET for the page and robots.txt; HEAD for links, or a one-byte ranged GET when a server refuses HEAD).

Limits

• Server HTML only. Content added by JavaScript is not seen; neither are lab metrics such as LCP or CLS.
• robots.txt is evaluated per RFC 9309 for Googlebot and * (wildcards, $, longest match, allow wins ties).
• At most 100 pages per run, 5 MB of HTML per page, 5 redirects.
• Requests to private, loopback, link-local and reserved addresses are refused, including after redirects.

Schemas and permissions

The input and every output item are described by JSON schemas that ship with the Actor (.actor/input_schema.json, .actor/dataset_schema.json). The Actor runs with Apify's limited permissions: it accesses only its own run's storage and the input you give it.