SEO Site Audit: 0-100 page score, broken links, fix hints

Crawl your website and get a technical SEO audit of every page: broken links, redirects, titles, meta descriptions, canonical, hreflang, noindex and more. One row per page with a 0-100 seoScore, an issues list and a plain-English fix for each issue, plus a site-wide SUMMARY with the average score and the worst pages. $1.50 per 1,000 pages.

🔎 What is SEO Site Audit Crawler?

It starts at your homepage, follows internal links and audits each HTML page. Then it checks every link target it found, so you see which links are broken and where. Each page gets issues with a code, a severity (error, warning, notice) and a message, e.g. TITLE_TOO_LONG: Title is 74 characters (over 60).

• 💯 SEO score and fixes per page: a transparent 0-100 score (see How the SEO score works) and one fix hint per issue, e.g. Add a <title> of 30-60 characters describing the page and its main keyword.
• 🔗 Broken links, really checked: internal targets (even beyond your page limit) and external links (HEAD, GET fallback), with status, anchor text and the page they are on.
• 🧭 Redirects in full: every hop (301/302/307/308), loops, temporary redirects, links to redirecting URLs.
• 🏷️ On-page checks: title, meta description, H1, canonical, meta robots / X-Robots-Tag, hreflang, Open Graph, structured data (JSON-LD, microdata, RDFa), word count, mixed content, viewport, lang, speed, size.
• 🤝 Polite: respects robots.txt (RFC 9309) and Crawl-delay, identifies itself as SEOSiteAuditBot.
• 💸 Fair pricing: only audited pages are charged; error pages, redirects and link checks are free.

Use cases: Analyze SEO & marketing (run website audits, get SEO & analytics data, extract website metadata & tech stacks) · Developer tools (monitor & alert on website changes).

Use cases

• SEO agencies: audit a client's site before a launch or redesign and hand over a spreadsheet.
• Broken link checker: 404s and dead external links, with the pages linking to them.
• Redirect checker after a migration.
• Meta tag audit: all titles, descriptions, H1s and canonicals in one table.
• Monitoring: weekly runs with alerts on new errors.
• Developers and QA: fail a release build if errorCount goes up.

Output fields

One row per crawled URL. Examples are from a real run on crawler-test.com (a site built for testing SEO crawlers).

Field	Description	Example
url / finalUrl	Crawled URL / URL after redirects	https://crawler-test.com/
status	HTTP status (null if the request failed)	200
redirectChain / redirectCount	Redirect hops	[] / 0
depth / foundOn	Clicks from start / page where first found	1 / https://crawler-test.com/
contentType	Content type	text/html
indexable	No noindex, canonical to itself or missing	true
title / titleLength	Title and length	Crawler Test Site / 17
metaDescription / metaDescriptionLength	Meta description and length	Default description... / 40
h1Count / h1 / h2Count	Headings	1 / ["Crawler Test Site"] / 0
canonical	Canonical link	https://crawler-test.com/mobile/separate_desktop
metaRobots / xRobotsTag	Robots meta and header	null
lang / hreflang	<html lang> / hreflang alternates	null / []
openGraph / twitterCard	Social tags	{} / null
wordCount	Visible words	1576
internalLinks / externalLinks / nofollowLinks	Link counts	411 / 3 / 3
brokenLinks / brokenLinkCount	Broken outgoing links (URL, status, error, anchor, internal)	see Output / 66
images / imagesMissingAlt / imagesMissingAltSamples	Images and missing alt text	0 / 0 / []
brokenImages	Broken images (if checked)	[]
structuredDataTypes / structuredDataFormats	Schema types, formats (JSON-LD, microdata, RDFa)	[]
responseTimeMs / sizeBytes	Speed and HTML size	328 / 42254
seoScore	0-100 page score from the issues (0 = URL did not load, null = not audited)	59
issues	{code, severity, message} list	see Output
fixes	One plain-English fix hint per issue, same order as issues	see Output
issueCount / errorCount / warningCount	Issue totals	10 / 1 / 3
error	Why the URL was not audited	null
crawledAt	Crawl time (UTC)	2026-10-05T03:24:23Z

🚀 How to use SEO Site Audit Crawler

1. Open the Store page and click Try for free.
2. Enter your homepage as the Start URL and set Max pages (default 20; raise it for a full site).
3. Click Start. The prefilled example (30 pages) takes about 1-2 minutes.
4. Download the rows as CSV, Excel or JSON and open the SUMMARY record.

Input guide

• Start URLs (startUrls): where the crawl starts; same-domain links are followed (www. counts as the same site), https:// is added when missing. Good: https://www.example.com/. Bad: a deep blog post alone, because the crawler only finds pages linked from where it starts. Private and local addresses are refused.
• Max pages (maxPages, default 20, max 10,000): pages, redirects and error pages all count, but only audited pages are charged. Links to pages beyond the limit are still checked for 404s.
• Max link depth (maxDepth, default 20): clicks from the start URL; 0 = start URL(s) only.
• Include subdomains (includeSubdomains): also crawl blog.example.com when you start on example.com.
• Check external links (checkExternalLinks, default on): HEAD requests, max 1 per second and 25 links per external site. Turn off for a faster run. Internal links are always checked.
• Check images (checkImages, default off): checks up to 200 image URLs per page for 404s. Missing alt text is always reported.
• URL patterns (includeUrlGlobs / excludeUrlGlobs): globs matched against the full URL and the path, applied to discovered links (not start URLs). Good: /blog/*, /tag/*, *?page=*. Bad: blog (no wildcard).
• Max link checks (maxLinkChecks, default 2000, max 50,000): cap on link targets not crawled as pages. Free, but slow servers make them slow.
• Parallel requests (maxConcurrency, default 5, max 20): a robots.txt Crawl-delay always wins. Keep it low for small servers.
• Request timeout (requestTimeoutSecs, default 20): slower pages are reported as Timed out; 429 and 5xx answers get one retry.
• Proxy (proxyConfiguration): usually not needed. Datacenter Apify Proxy only; residential proxies make the run fail.

Other limits: HTML up to 5 MB, 10 redirects per chain; file links (PDF, images, ZIP...) are not crawled as pages.

Full example input:

{
  "startUrls": [{ "url": "https://crawler-test.com/" }],
  "maxPages": 30,
  "maxDepth": 20,
  "includeSubdomains": false,
  "checkExternalLinks": true,
  "checkImages": false,
  "includeUrlGlobs": [],
  "excludeUrlGlobs": ["/tag/*", "*?page=*"],
  "maxLinkChecks": 2000,
  "maxConcurrency": 5,
  "requestTimeoutSecs": 20,
  "proxyConfiguration": { "useApifyProxy": false }
}

📦 Output

Homepage row from a real run on crawler-test.com, 2026-10-05 (brokenLinks, issues and fixes shortened; 1 error, 3 warnings and 6 notices give 100 - 20 - 15 - 6 = 59):

{
  "url": "https://crawler-test.com/",
  "finalUrl": "https://crawler-test.com/",
  "status": 200,
  "redirectChain": [],
  "redirectCount": 0,
  "depth": 0,
  "foundOn": null,
  "contentType": "text/html",
  "indexable": true,
  "title": "Crawler Test Site",
  "titleLength": 17,
  "metaDescription": "Default description XIbwNE7SSUJciq0/Jyty",
  "metaDescriptionLength": 40,
  "h1Count": 1,
  "h1": ["Crawler Test Site"],
  "h2Count": 0,
  "canonical": null,
  "metaRobots": null,
  "xRobotsTag": null,
  "lang": null,
  "hreflang": [],
  "openGraph": {},
  "twitterCard": null,
  "wordCount": 1576,
  "internalLinks": 411,
  "externalLinks": 3,
  "nofollowLinks": 3,
  "brokenLinks": [
    {"url": "http://www.søkbar.no/", "status": null, "error": "Domain not found (DNS lookup failed)",
     "anchor": "Foreign Character Domain", "internal": false},
    {"url": "https://crawler-test.com/redirects/redirect_to_404", "status": 404, "error": null,
     "anchor": "Redirect To 404 Http Status", "internal": true}
  ],
  "brokenLinkCount": 66,
  "images": 0,
  "imagesMissingAlt": 0,
  "imagesMissingAltSamples": [],
  "brokenImages": [],
  "structuredDataTypes": [],
  "structuredDataFormats": [],
  "responseTimeMs": 328,
  "sizeBytes": 42254,
  "seoScore": 59,
  "issues": [
    {"code": "BROKEN_INTERNAL_LINKS", "severity": "error", "message": "64 broken internal link(s)"},
    {"code": "MIXED_CONTENT", "severity": "warning", "message": "2 insecure http:// resource(s) on an https page"},
    {"code": "TITLE_TOO_SHORT", "severity": "notice", "message": "Title is only 17 characters (under 30)"},
    {"code": "LINKS_TO_REDIRECTS", "severity": "notice", "message": "10 internal link(s) point to redirecting URLs"}
  ],
  "fixes": [
    "Fix or remove the broken internal links listed in brokenLinks (or redirect their targets).",
    "Load all scripts, styles and images over https:// instead of http://.",
    "Expand the <title> to 30-60 characters with a descriptive keyword phrase.",
    "Update internal links to point directly at the final URL instead of a redirecting one."
  ],
  "issueCount": 10,
  "errorCount": 1,
  "warningCount": 3,
  "error": null,
  "crawledAt": "2026-10-05T03:24:23Z"
}

Dataset views

Table views in Console (API: ?view=<name>):

• Scores and issues per page (issues): SEO score, status, indexable, issue counts, broken links, issue list, fix hints, error.
• Titles and meta tags (meta): title, meta description, H1, canonical, robots meta, word count, images without alt.
• Links and redirects (links): redirect chain, final URL, depth, link counts, broken links, response time.

Key-value store records

• SUMMARY: site-wide view. averageSeoScore (over all scored rows), seoScoreBands (good 90-100, needsWork 50-89, poor 0-49), worstPages (10 lowest scores with their top issues and fixes), issue counts by code and severity, status codes, top 50 broken link targets with how many pages link to them, duplicate titles, meta descriptions and content, slowest pages, robots.txt info (crawl-delay, sitemaps) and URLs blocked by robots.txt. In the run above: 30 pages, 27 indexable, 66 broken link targets, 7 missing meta descriptions.
• RUN_SUMMARY: run stats (pages crawled, audited and charged, requests, link checks, speed; the error if the input was invalid).

Export

Download as JSON, CSV, Excel, XML, RSS or HTML table, or via the API (/v2/datasets/{datasetId}/items?format=csv&view=meta). Duplicates are checked only between indexable pages (no noindex, canonical pointing to itself or missing). Links that answer 401, 403, 429 or 999 (common on social networks that refuse bots) are not counted as broken.

💯 How the SEO score works

Every row gets seoScore = 100 minus a fixed deduction per issue, by severity, floored at 0:

Severity	Deduction	Examples
error	-20	TITLE_MISSING, CANONICAL_BROKEN, BROKEN_INTERNAL_LINKS
warning	-5	META_DESCRIPTION_MISSING, H1_MISSING, TITLE_TOO_LONG, IMAGES_MISSING_ALT, NOINDEX, REDIRECT_CHAIN, DUPLICATE_TITLE
notice	-1	TITLE_TOO_SHORT, H1_MULTIPLE, CANONICAL_MISSING, OG_TAGS_MISSING, LANG_MISSING

• Each issue code counts once per page, however many links or images it covers (e.g. 64 broken internal links = one -20).
• A URL that did not load (HTTP_4XX, HTTP_5XX, FETCH_FAILED, REDIRECT_LOOP, TOO_MANY_REDIRECTS) scores 0.
• Rows that were not audited (blocked by robots.txt, non-HTML files, redirects to a page that has its own row) get null and are left out of the averages.
• Rough reading: 90-100 good, 50-89 needs work, 0-49 poor. Recompute it yourself from issues if you want different weights.

fixes[i] is the fix for issues[i], e.g. META_DESCRIPTION_MISSING → Add a meta description of 70-160 characters summarising the page.

Issue codes

• Errors: HTTP_4XX, HTTP_5XX, FETCH_FAILED, REDIRECT_LOOP, TOO_MANY_REDIRECTS, TITLE_MISSING, CANONICAL_BROKEN, BROKEN_INTERNAL_LINKS
• Warnings: REDIRECT_CHAIN, TITLE_TOO_LONG, TITLE_MULTIPLE, META_DESCRIPTION_MISSING, META_DESCRIPTION_MULTIPLE, H1_MISSING, CANONICAL_MULTIPLE, NOINDEX, BROKEN_EXTERNAL_LINKS, BROKEN_IMAGES, IMAGES_MISSING_ALT, SLOW_RESPONSE (> 2 s), MIXED_CONTENT, HTTP_NOT_HTTPS, VIEWPORT_MISSING, HREFLANG_INVALID, HREFLANG_RELATIVE, STRUCTURED_DATA_INVALID, DUPLICATE_TITLE, DUPLICATE_META_DESCRIPTION, DUPLICATE_CONTENT
• Notices: TEMPORARY_REDIRECT, BLOCKED_BY_ROBOTS, NOT_HTML, TITLE_TOO_SHORT (< 30), META_DESCRIPTION_TOO_LONG (> 160), META_DESCRIPTION_TOO_SHORT (< 70), H1_MULTIPLE, CANONICAL_MISSING, CANONICAL_TO_OTHER_URL, NOFOLLOW_PAGE, LINKS_TO_REDIRECTS, LOW_WORD_COUNT (< 200), LARGE_PAGE (> 1 MB), LANG_MISSING, OG_TAGS_MISSING, HREFLANG_NO_SELF, META_REFRESH, URL_TOO_LONG (> 115)

💵 Pricing: $1.50 per 1,000 pages

You pay once per page that was audited (event page: an HTML page that answered 2xx), plus Apify's $0.00005 per-run start fee. Platform compute is included. Error pages (404, 500), redirects, robots.txt-blocked URLs and all link checks are free.

Run	Audited pages	Cost
Prefilled example	30	$0.045
Small business site	100	$0.15
Medium site	1,000	$1.50
Large site (maximum per run)	10,000	$15.00

Set Max charge per run to cap spending: the crawl then stops at the number of pages your limit covers. Other SEO audit actors on the Store charge $2-40 per 1,000 pages.

🔌 Use it via API

JavaScript (npm install apify-client):

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_API_TOKEN' });
const run = await client.actor('kadi_bence/seo-site-audit').call({
  startUrls: [{ url: 'https://www.example.com/' }],
  maxPages: 500,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();

Python (pip install apify-client):

from apify_client import ApifyClient

client = ApifyClient("YOUR_API_TOKEN")
run = client.actor("kadi_bence/seo-site-audit").call(run_input={
    "startUrls": [{"url": "https://www.example.com/"}],
    "maxPages": 500,
})
for page in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(page["url"], page["seoScore"], page["fixes"][:3])
summary = client.key_value_store(run["defaultKeyValueStoreId"]).get_record("SUMMARY")

cURL (waits for the run and returns the rows):

curl -X POST "https://api.apify.com/v2/acts/kadi_bence~seo-site-audit/run-sync-get-dataset-items?token=YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"startUrls": [{"url": "https://www.example.com/"}], "maxPages": 100}'

Apify CLI:

apify call kadi_bence/seo-site-audit --input '{"startUrls": [{"url": "https://www.example.com/"}], "maxPages": 100}'
apify call kadi_bence/seo-site-audit -f input.json

MCP for AI agents (Claude, Cursor, VS Code and other MCP clients):

{"mcpServers": {"apify": {"url": "https://mcp.apify.com/?actors=kadi_bence/seo-site-audit"}}}

🔁 Integrations & scheduling

• Schedules: run audits on a cron, e.g. 0 6 * * 1 = every Monday at 06:00.
• Webhooks: on "run succeeded", Apify calls your URL with the run ID.
• Zapier, Make, n8n, Google Sheets, Slack, email: pass results on with Apify integrations.

Recipes:

1. Weekly SEO audit emailed to a client. Save a task with the client's homepage and maxPages: 1000, schedule it every Monday, and let a Make or Zapier scenario email the Scores and issues per page view as CSV with the SUMMARY totals. About $1.50 a week.
2. Migration check. Run with checkExternalLinks: false, then filter the issues for REDIRECT_CHAIN, REDIRECT_LOOP and LINKS_TO_REDIRECTS.
3. Broken link alert. A webhook calls your script; if SUMMARY brokenLinkTargets is above zero, post topBrokenLinks to Slack.

🧰 Related tools by the same developer

Website due diligence:

• Website Tech Stack Detector: CMS, ecommerce, analytics and frameworks from 7,600+ fingerprints. $1.50 / 1,000 websites.
• Website Screenshot & PDF API: full-page and mobile PNG, JPEG, WebP and PDF captures. $2.50 / 1,000 screenshots.
• Bulk WHOIS & RDAP Domain Lookup: registrar, age, expiry, DNS/MX/SPF/DMARC. $1.50 / 1,000 domains.

Content pipelines:

• Document to Markdown: PDF, Word, PowerPoint, Excel and OCR to Markdown and RAG chunks. $2 / 1,000 documents.
• Bulk Image Downloader: all images and files from pages into a ZIP. $0.80 / 1,000 files.

❓ FAQ

How does it work? Plain HTTP, no browser: it crawls internal links breadth-first, then checks all link targets found, compares pages for duplicates and writes the SUMMARY.

Which sites can I crawl? Your own sites, or sites you have permission to audit. It stays on the domain(s) of your start URLs; other sites only get single HEAD requests to check your outgoing links.

Does it respect robots.txt? Always: RFC 9309 (group for SEOSiteAuditBot, else *; longest match wins; * and $ wildcards) plus Crawl-delay. Disallowed URLs are never requested and are listed in the SUMMARY. If robots.txt answers 5xx or times out, the site is treated as "do not crawl". Pages with nofollow are audited, but their links are not followed. To audit blocked pages, add an Allow rule for User-agent: SEOSiteAuditBot.

How fast is it? The crawl ran at about 700 pages per minute in our tests (5 parallel requests, server answering in about 0.4 s). Link checks come after; on a slow server 430 link targets took about 80 s in our 200-page test. Large sites need a longer run timeout; the Actor stops before it and still saves results.

What are the limits? 10,000 pages and 50,000 link checks per run, HTML up to 5 MB. No JavaScript rendering: single-page apps show what the server sends, as many search engines see it on the first pass. Sitemap-only and URL-list modes are on the roadmap.

My site blocks the crawler. Do I need a proxy? Usually not. Allow the user agent SEOSiteAuditBot in your firewall or CDN, or try the datacenter Apify Proxy.

Why is an external link OK here but broken in my browser? Some sites (LinkedIn, Instagram, some shops) answer 401/403/429/999 to every bot, so those are not reported. 404, 410, other 4xx, 5xx, dead domains, refused connections, timeouts and redirect loops are.

Does it collect personal data? Is it legal? It reads SEO tags, not personal data. E-mails and phone numbers in titles, descriptions, headings or anchors become [email removed] / [phone removed]; mailto:/tel: links are ignored. Getting permission to crawl is your responsibility; this is not legal advice.

What does a big job cost? 10,000 pages cost $15. Use Max charge per run as a hard cap.

What if it fails? Failed pages are free rows with status, an error text and an issue such as FETCH_FAILED or HTTP_4XX. Invalid input stops the run before any page is charged. Something wrong? Open an issue on the Issues tab with the URL. Fixes usually land within 24-48 hours.

Can I use it from Python, integrations or an AI agent? Yes, see Use it via API and Integrations above. Version history is in the Changelog.

📝 Changelog

• 1.0 (2026-10-05): First release. Full-site crawl with 47 issue codes, broken internal and external links (HEAD with GET fallback), redirect chains, duplicates, hreflang and structured data checks, robots.txt per RFC 9309 with Crawl-delay, SUMMARY record, pay only for audited pages.
• 2026-10-05: Store page rewritten.
• 1.1 (2026-10-05): 0-100 seoScore per page (transparent severity weights), a plain-English fix hint per issue (fixes), and averageSeoScore, seoScoreBands and worstPages in the SUMMARY.

🤝 Need a custom version?

I build custom scrapers, scheduled data feeds and integrations (CSV, Excel, Google Sheets, API, MCP for AI agents). Tell me the sites and fields you need:

• Email: bence.kadi@gmail.com
• Apify profile: https://apify.com/kadi_bence

Related Actors

More low-cost Actors by the same developer, built on official APIs and public data:

• PageSpeed Insights Bulk Checker — Core Web Vitals and Lighthouse scores in bulk
• Website Tech Stack Detector — CMS, ecommerce, analytics and frameworks of any website
• AI Visibility Tracker — how ChatGPT, Perplexity, Gemini and Claude mention your brand
• Bulk WHOIS & RDAP Domain Lookup — registrar, dates and DNS for many domains
• Website Screenshot & PDF API — full-page screenshots and PDFs of any URL
• Bulk Image Downloader — download all images from URLs as a ZIP
• Document to Markdown — PDF, Word, PowerPoint and Excel to clean Markdown, with OCR

All my Actors: apify.com/kadi_bence