Technical SEO Audit & Monitor: Broken Links, Redirects

Audit a website for technical SEO issues and get one row per page: broken links, redirects and redirect chains, noindex, canonical problems, missing or duplicate titles and meta descriptions, H1, hreflang, structured data and sitemap coverage, with a 0-100 score per page.

Run it on a schedule with change tracking and it becomes an SEO monitor: each run tells you which pages broke, got de-indexed, lost their title or left the sitemap since the previous audit — and only those pages are saved and charged.

Try it with the prefilled input: 20 pages of python.org, results in about half a minute.

What does this technical SEO audit tool do?

It crawls a website the way a search engine does, starting from the URL you give it and following internal links, then reads the sitemap to find pages that no link points to. Every URL is requested without following redirects, so a 301 is reported on the URL that answers it, not hidden behind the final page.

Each page gets:

• its HTTP status, redirect target, response time and HTML size;
• its indexability, and the reason when it is not indexable (noindex, canonical to another page, redirect, error);
• the SEO tags that matter: title, meta description, H1, canonical, robots meta and X-Robots-Tag, lang, hreflang alternates, JSON-LD types;
• its links: internal and external counts, the broken internal pages it links to, the links that redirect, and how many crawled pages link to it;
• a list of issues with a severity (error, warning, notice) and a plain-English message, and a score.

Site-wide checks run once the crawl is done: duplicate titles and descriptions, pages missing from the sitemap, sitemap entries that are broken or not indexable, and orphan pages (in the sitemap, linked from nowhere).

It is plain HTTP, no browser: fast and cheap. The trade-off is stated below in the FAQ.

Why audit and monitor technical SEO?

• Catch regressions before rankings drop. A deploy that adds noindex to a template, a CMS migration that breaks 200 URLs, a canonical that suddenly points to the homepage: schedule the Actor daily and get only those pages.
• Agency and freelance reporting. Audit every client site on a schedule, export to Google Sheets or Excel, and show the issues fixed since last month (resolvedIssues).
• Pre- and post-migration checks. Compare the site before and after a redesign or a domain move.
• Content hygiene at scale. Find every page with a missing meta description or a duplicate title in one run, instead of clicking through a CMS.
• Feed your own tools. A clean JSON row per URL, ready for a dashboard, a data warehouse or an AI agent.

How to run a technical SEO audit

1. Enter the website in Websites to audit (https://www.example.com/ or just example.com). Several websites can go in one run.
2. Set Max pages per site. Start with the default to check the output, then raise it above the size of the site.
3. Click Start, then open the Issues view of the Output tab.
4. To monitor the site, open Schedules, run the Actor daily or weekly, and switch on Save only changes.

What happens during a run

1. robots.txt is read first. URLs it disallows are never requested, and never charged.
2. The start page is fetched, then every internal link is followed, page by page, up to your limit. Redirects are recorded, not followed blindly: each URL keeps its own status code.
3. The sitemap is read and its URLs that no link points to are audited last, which is how orphan pages are found.
4. Site-wide checks run once the crawl is done — duplicate titles and descriptions, broken internal links, sitemap coverage — because each of them needs every page at once.
5. Results are saved, one row per URL, and compared with the previous audit when tracking is on.

Files that are not web pages (PDF, images, archives) are never requested, and a response that is not HTML is recorded from its headers alone, so you are not charged for downloading them.

Input

Field	Default	What it does
startUrls	–	Websites to audit, one per line (URL or domain).
maxPagesPerSite	20	Pages audited per website (up to 2,000).
maxItems	100	Pages audited in the whole run.
useSitemap	true	Read the sitemap to find orphan pages and check sitemap coverage.
includeSubdomains	false	Also crawl blog.example.com when auditing example.com.
excludeUrlPatterns	–	Skip URLs containing any of these texts (/tag/, ?replytocom=).
trackChanges	false	Mark each page new, changed, unchanged or removed.
onlyChanges	false	Save only new, changed and removed pages (implies trackChanges).
stateStoreName	technical-seo-audit-state	Named key-value store holding the previous audit.
maxConcurrency	4	Pages requested in parallel.
maxRequestsPerMinute	120	Crawl rate limit.
proxyConfiguration	no proxy	Only for sites that block datacenter IPs.

{
    "startUrls": ["https://www.example.com/"],
    "maxPagesPerSite": 500,
    "maxItems": 500,
    "onlyChanges": true,
    "excludeUrlPatterns": ["/tag/", "?replytocom="]
}

Output

One item per audited URL. You can download the dataset as JSON, CSV, Excel, XML or HTML, or read it through the API. A shortened example:

{
    "url": "https://www.example.com/pricing/",
    "siteUrl": "https://www.example.com/",
    "statusCode": 200,
    "redirectUrl": null,
    "responseTimeMs": 182,
    "depth": 1,
    "inlinkCount": 14,
    "inSitemap": true,
    "indexable": true,
    "nonIndexableReason": null,
    "title": "Pricing | Example",
    "titleLength": 17,
    "metaDescription": null,
    "h1": "Simple pricing",
    "h1Count": 1,
    "canonicalUrl": "https://www.example.com/pricing/",
    "noindex": false,
    "hreflang": [{ "lang": "fr", "url": "https://www.example.com/fr/tarifs/" }],
    "jsonLdTypes": ["Product"],
    "internalLinkCount": 48,
    "externalLinkCount": 3,
    "brokenLinks": [{ "url": "https://www.example.com/old-plan/", "statusCode": 404 }],
    "redirectingLinkCount": 2,
    "issues": [
        { "code": "broken_links", "severity": "error", "message": "The page links to 1 broken page(s) of the site." },
        { "code": "links_to_redirect", "severity": "warning", "message": "The page links to 2 URL(s) that redirect." },
        { "code": "missing_meta_description", "severity": "warning", "message": "The page has no meta description." },
        { "code": "title_too_short", "severity": "notice", "message": "The title is 17 characters (under 30)." }
    ],
    "errorCount": 1,
    "warningCount": 2,
    "noticeCount": 1,
    "score": 69,
    "changeStatus": "changed",
    "changedFields": ["metaDescription"],
    "previousValues": { "metaDescription": "Plans for every team." },
    "newIssues": ["missing_meta_description"],
    "resolvedIssues": [],
    "previousScrapedAt": "2026-09-16T06:00:04.112Z",
    "scrapedAt": "2026-09-17T06:00:03.871Z"
}

Fields you get for every page

Field	What it tells you
statusCode, redirectUrl, fetchError	What the URL answers, and where it redirects
indexable, nonIndexableReason	Whether a search engine can index it, and why not
title, titleLength, metaDescription, h1, h1Count	The tags that show up in search results
canonicalUrl, metaRobots, xRobotsTag, noindex	Indexing directives, from the page and from the headers
lang, hreflang, jsonLdTypes	Language targeting and structured data
brokenLinks, redirectingLinkCount, internalLinkCount	Which links on the page are broken or redirect
inlinkCount, depth, inSitemap	How the page is reached, and whether the sitemap lists it
responseTimeMs, htmlSizeBytes, wordCount	Speed and weight, and how much text the page has
issues, errorCount, warningCount, score	Every problem found, with a severity and a plain message
changeStatus, changedFields, newIssues, previousValues	What changed since the previous audit

The run summary

Besides the per-page rows, each run writes a SUMMARY record to the key-value store, with one entry per website: pages audited, average score, issue counts, the most common issue codes, whether the crawl covered the whole site, and why a website could not be audited when that happens.

{
    "sitesRequested": 1,
    "pagesSaved": 500,
    "sites": [
        {
            "siteUrl": "https://www.example.com/",
            "status": "ok",
            "pagesAudited": 500,
            "complete": true,
            "averageScore": 88,
            "issueCounts": { "error": 4, "warning": 121, "notice": 260 },
            "topIssues": [
                { "code": "meta_description_too_short", "pages": 96 },
                { "code": "duplicate_title", "pages": 41 },
                { "code": "broken_links", "pages": 4 }
            ]
        }
    ],
    "budgetReached": false
}

Issues checked

Severity	Codes
Error	http_4xx, http_5xx, fetch_failed, broken_links, redirect_to_error, redirect_without_location, missing_title, canonical_to_error
Warning	redirect_chain, links_to_redirect, missing_meta_description, missing_h1, duplicate_title, duplicate_meta_description, canonical_to_redirect, sitemap_non_indexable, orphan_page, invalid_json_ld, invalid_hreflang, hreflang_on_non_indexable, mixed_content, missing_viewport, slow_response
Notice	redirect, title_too_long (> 60), title_too_short (< 30), meta_description_too_long (> 160), meta_description_too_short (< 70), multiple_h1, missing_canonical, canonicalized, noindex, nofollow_page, not_in_sitemap, missing_lang, low_word_count (< 100), large_html (> 2 MB)

The score starts at 100 and loses 20 points per error, 5 per warning and 1 per notice, down to 0. Title, description and heading checks only apply to indexable pages.

Monitoring: tracking changes between audits

With trackChanges (or onlyChanges), each page is compared with the previous run of the same website:

• new — not seen in the previous audit;
• changed — status, redirect, indexability, noindex, canonical, title, meta description, H1 or sitemap presence changed, or issues appeared or were fixed (changedFields, previousValues, newIssues, resolvedIssues);
• unchanged — nothing moved (not saved with onlyChanges);
• removed — audited last time, no longer reached by the crawl (unlinked, deleted, or now excluded).

Removed pages are only reported when the crawl covered the whole site: if the page limit cut it short, the Actor says so in the summary and never reports pages as removed. Set maxPagesPerSite above the size of the site for monitoring. The first tracked run marks every page as new.

Use it from the API, integrations or an AI agent

curl -X POST "https://api.apify.com/v2/acts/succinct_glider~technical-seo-audit/run-sync-get-dataset-items?token=<YOUR_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"startUrls": ["example.com"], "maxPagesPerSite": 50}'

Send the results to Google Sheets, Slack, a webhook, Make, Zapier or n8n with Apify integrations, or let an AI agent call the Actor through the Apify MCP server.

Pricing

$5 per 1,000 audited pages ($0.005 per page). No start fee.

• Only saved pages are charged. With onlyChanges, unchanged pages are free, so monitoring a stable 500-page site costs a few cents per run.
• A site that cannot be audited (blocked, down, disallowed by robots.txt) is reported in the summary and costs nothing.
• Set a maximum cost per run in the run options: the Actor stops cleanly when it is reached.

Tips

• Exclude the noise with excludeUrlPatterns: tag archives, search results, calendar pages, ?sort= parameters. Excluded URLs are never requested.
• Keep the rate gentle on small servers: maxConcurrency 2 and maxRequestsPerMinute 60.
• Use one stateStoreName per monitoring job if you run several schedules on the same website with different settings.
• Start from a section (https://example.com/blog/): the crawl starts there but still follows every link of the site.

FAQ

Does it render JavaScript? No. It reads the HTML the server sends, like most SEO crawlers in their default mode. Pages whose content, titles or links are injected by JavaScript after load will look emptier than they are in a browser. Single-page apps without server-side rendering are not a good fit.

How large a website can it audit? Up to 2,000 pages per website per run, which keeps a run within the default 1 GB of memory. For larger websites, audit sections separately with excludeUrlPatterns, and open an issue if you need more.

Does it check external links? Not yet: it counts them, but only internal pages are requested. External link checking is planned.

Can it audit a staging site behind a login or a password? No. It only reads pages a visitor can open without signing in, and it never fills in credentials. Put the staging site behind an IP allowlist rather than a password if you want to audit it.

How long does a run take, and what does it cost in compute? About 120 pages per minute with the default rate limit, so a 1,000-page audit takes around 10 minutes. On a 1,000-page run measured on the platform, compute was about $0.08 — and platform usage is included in the price per page, so it is not billed to you on top.

Does it respect robots.txt? Always, and it cannot be turned off. Pages that robots.txt disallows are skipped and not charged. The crawler identifies itself as TechnicalSeoAuditBot in its User-Agent, and it only follows links within the website you entered.

Is it legal to audit a website? Audit your own websites or your clients' with their permission. The Actor only requests public pages, respects robots.txt, crawls at a modest rate, and does not store page content — only SEO tags, statuses and counts. You are responsible for the websites you choose to audit and for complying with their terms.

A website came back with no results. The run summary tells you why: unreachable (the start page answered an error or blocked the request — try a proxy), disallowed (robots.txt), or redirected_away (the URL redirects to another domain — audit that one instead). None of these fail the run, and none is charged.

Why is my page marked as not in the sitemap when it is? URLs are compared exactly: https://example.com/page and https://example.com/page/ are different URLs for a search engine. If the sitemap is very large (over 50,000 URLs), sitemap coverage is not checked at all rather than reported wrongly.

Support

Found a false positive, or a check you would like to see? Open an issue on the Issues tab with the page URL and the issue code, and I will look at it.