Broken Link Checker: Find Dead Links, 404s & Redirects

Check your website for broken links and dead links: a broken link checker for site owners, SEO audits and content migrations.

Give it your website (or its sitemap) and get a list of every broken link, error page and redirect chain, with the page where each link appears and its anchor text. Each unique URL is checked once, internal links always and external links when you switch them on.

Quick start for AI agents and API users

• What it does: Find broken links (404, 5xx, DNS, timeouts) and redirect chains on one website.
• When to use it: For site audits after a redesign.
• Example input (the form's prefill; it finishes in under a minute):

{
  "startUrls": [
    "https://crawlee.dev"
  ],
  "maxPages": 2,
  "maxUrlsChecked": 15,
  "requestIntervalSecs": 0.5
}

• Example output: one real dataset row from run tI03Q30d6yOLLf0Rt (2026-10-23), with long fields shortened and empty fields left out. Fields not shown are listed in the output table below.

{
  "url": "https://crawlee.dev/",
  "finalUrl": "https://crawlee.dev/",
  "status": 200,
  "category": "ok",
  "redirectCount": 0,
  "contentType": "text/html",
  "responseTimeMs": 20,
  "method": "GET",
  "isInternal": true,
  "occurrences": 3,
  "foundOn": [
    {
      "pageUrl": "https://crawlee.dev/",
      "anchorText": "Crawlee",
      "tag": "a"
    },
    {
      "pageUrl": "https://crawlee.dev/js",
      "anchorText": "Crawlee",
      "tag": "a"
    }
  ],
  "firstFoundOn": "https://crawlee.dev/",
  "anchorText": "Crawlee",
  "depth": 0,
  "site": "crawlee.dev",
  "checkedAt": "2026-10-23T19:10:50Z"
}

• Cost (pay per event, Free-plan prices): Actor start $0.005; Checked URL $0.0002. Failed items are not charged.
• From an AI agent: with the Apify MCP server, call call-actor with actor phaux/broken-link-checker, or load it as its own tool with https://mcp.apify.com?tools=phaux/broken-link-checker. Read results with get-dataset-items.

What does Broken Link Checker do?

• Crawls your site from a start URL, following internal links up to maxDepth and maxPages, or starts from the pages listed in a sitemap (https://example.com/sitemap.xml, sitemap indexes and .xml.gz included).
• Checks every unique link once: the final HTTP status code, every redirect hop (301, 302, 307, 308) with its status, and network errors (DNS, connection refused, timeout, too many redirects).
• Shows where each link is used: up to 10 source pages per URL with anchor text (or image alt text), plus how many pages link to it.
• Optional checks: external links, and images, scripts, stylesheets and iframes (checkResources).
• Sorts broken links first and saves a SUMMARY with the broken links and their source pages.

Why use it?

• Fix 404s before your users and search engines find them, after a redesign, migration or content cleanup.
• Find redirect chains (A → B → C) that waste crawl budget and slow pages down; redirectChain lists every hop.
• Pay only per checked URL; links that can't be checked (robots.txt, caps, time budget) are listed for free.
• Safe defaults: 50 pages and 1,000 checked URLs per run unless you raise them, and a time guard that saves results before the run timeout.

Use cases: SEO audits, website QA before a launch, link rot monitoring on blogs and documentation, migration checks (old URLs should redirect once, not three times), and agency site audits.

How to use it

1. Put your site's start page (https://example.com) or sitemap URL in Start URL or sitemap. Empty input runs a small example on crawlee.dev.
2. Set Max pages to crawl and Max URLs checked (the second one caps your cost).
3. Switch on Check external links to also verify outbound links.
4. Run it and open the All checked URLs view; filter by category = broken. Export as CSV, Excel or JSON.

Input

{
  "startUrls": ["https://crawlee.dev"],
  "maxPages": 200,
  "maxUrlsChecked": 2000,
  "maxDepth": 3,
  "checkExternalLinks": true,
  "checkResources": false,
  "excludeUrlPatterns": ["/tag/", "\\?replytocom="],
  "requestIntervalSecs": 1
}

Field	Type	Default	Description
startUrls	array	—	Your website's start page (e.g. https://example.com) or a sitemap URL (e.g. https://example.com/sitemap.xml). Several URLs are allowed if they are on the same website. Empty = a small example run on crawlee.dev.
maxPages	integer	50	How many internal HTML pages are opened to collect links (1–10,000). Links on those pages are all checked, so the number of checked URLs is usually larger.
maxUrlsChecked	integer	1000	Upper limit of checked URLs (= results you pay for). 0 = up to 100,000. URLs over the limit are listed in SKIPPED_LINKS for free.
maxDepth	integer	3	Link depth from the start page (0 = only check the links on the start pages).
checkExternalLinks	boolean	false	Also check links that point to other websites (one request each, never crawled further).
checkResources	boolean	false	Also check img, script, iframe, video/audio source and stylesheet/icon URLs, not only a/area links.
useSitemap	boolean	false	Add the pages from the site's sitemaps (found via robots.txt or /sitemap.xml) to the crawl, up to Max pages. Useful for finding orphan pages.
includeUrlPatterns	array	—	Regular expressions; when set, only internal pages whose URL matches one of them are crawled for links (links found on them are still all checked). Example: /blog/
excludeUrlPatterns	array	—	Regular expressions; matching URLs are neither crawled nor checked. Example: /logout|?replytocom=
requestIntervalSecs	number	1	Politeness delay per host (0.5–30 s). A robots.txt Crawl-delay overrides it when longer.
maxConcurrency	integer	5	Requests in flight across different hosts (1–20). Requests to the same host are always one at a time.
timeBudgetSecs	integer	0	Stop crawling and checking after this many seconds and save the results so far (0 = only the run timeout guard, which always applies).

Output example

One dataset item per checked URL:

{
  "url": "https://example.com/old-pricing",
  "category": "broken",
  "status": 404,
  "finalUrl": "https://example.com/pricing-2024",
  "redirectCount": 1,
  "redirectChain": [{"url": "https://example.com/old-pricing", "status": 301}],
  "errorCode": null,
  "isInternal": true,
  "firstFoundOn": "https://example.com/blog/launch",
  "anchorText": "see our pricing",
  "occurrences": 3,
  "foundOn": [{"pageUrl": "https://example.com/blog/launch", "anchorText": "see our pricing", "tag": "a"}],
  "contentType": "text/html",
  "responseTimeMs": 182,
  "method": "HEAD",
  "depth": 2,
  "site": "example.com",
  "checkedAt": "2026-10-20T21:40:00Z"
}

Output fields

Field	Type	Description
url	string	Checked URL (normalised, without #fragment).
category	string	broken (4xx/5xx or no response), redirect (ended OK after one or more redirects) or ok.
status	integer or null	Final HTTP status code; null when there was no HTTP response (DNS, connection or timeout error).
finalUrl	string or null	URL after following redirects.
redirectCount	integer	Number of redirect hops.
redirectChain	array	Each hop as {url, status}, in order, before finalUrl.
errorCode	string or null	dns_error, network_error, timeout, too_many_redirects, blocked_redirect or http_error when the check failed without a usable response.
error	string or null	Human-readable error.
isInternal	boolean	true when the URL is on the audited website (www. ignored).
firstFoundOn	string or null	First page where the link was found (null for start or sitemap URLs).
anchorText	string or null	Link text (or image alt) on firstFoundOn.
occurrences	integer	How many crawled pages link to this URL.
foundOn	array	Up to 10 source pages as {pageUrl, anchorText, tag}.
contentType	string or null	Content-Type of the final response.
responseTimeMs	integer or null	Time of the final request in milliseconds.
method	string	HEAD or GET (GET is used for crawled pages and when HEAD fails).
depth	integer	Link depth from the start page.
site	string	Audited website.
checkedAt	string	UTC time of the run.

The key-value store also gets SUMMARY (counts and the broken links with their source pages), SKIPPED_LINKS (URLs not checked, with a code: robots_disallowed, host_stopped, rate_limited, blocked_host, max_urls, time_budget) and RUN_STATS.

How much does it cost?

Pay per event, the same on every Apify plan. Apify platform usage is included.

Event	Price
Actor start (once per run)	$0.005
Checked URL (per result saved)	$0.0002

Example: 1,000 checked URLs cost $0.205; the small example run (17 URLs) costs about $0.0084. Skipped links and bad-input runs that stop before crawling are never charged per URL. Use Max URLs checked to cap any run.

Use with AI agents, MCP and the API

• MCP: add https://mcp.apify.com?tools=phaux/broken-link-checker to an MCP client and ask, for example, "find broken links on example.com".
• API: POST https://api.apify.com/v2/acts/phaux~broken-link-checker/run-sync-get-dataset-items with {"startUrls": ["https://example.com"], "maxUrlsChecked": 300} returns the results directly.
• Every input and output field is described in the input and dataset schemas.

Data source and compliance

• Meant for your own websites (or sites you are allowed to audit). Only public http(s) URLs are fetched; private and internal addresses are refused on every redirect hop.
• robots.txt is honoured (RFC 9309) for the user agent PhauxLinkCheckBot, including Crawl-delay. Disallowed links are skipped and listed in SKIPPED_LINKS.
• Polite pacing: requests to one host are sequential and at least requestIntervalSecs apart (default 1 s). After an HTTP 429 a host is not contacted again in that run; if your own site answers 429, the crawl stops and the results so far are saved.
• It identifies itself honestly (Mozilla/5.0 (compatible; PhauxLinkCheckBot/0.1; +https://apify.com/phaux/broken-link-checker)), uses no proxies and no IP rotation, never logs in and doesn't get around bot protection.
• External links are checked with one HEAD request (GET when HEAD fails) and are never crawled further.
• Anchor texts are stored with email addresses and phone numbers removed.

Limits

• One website per run (several start URLs are fine if they share the host; www. is ignored). Subdomains count as external.
• JavaScript-rendered links are not seen: the Actor reads the HTML the server sends.
• Up to 10,000 pages crawled and 100,000 URLs checked per run. At the default 1 request per second, 1,000 internal URLs take roughly 17 minutes.
• Some servers answer 403 or 429 to automated checks even when the page works in a browser; those show as broken or skipped with the real status code.

FAQ

• Why is a link "broken" when it works in my browser? The server returned an error status to the checker (often 403 from bot protection, or 405 to HEAD; HEAD failures are retried with GET). The status field shows exactly what the server answered.
• How do I see only broken links? Filter the dataset by category = broken, or open the SUMMARY record.
• What are redirect chains? redirectChain lists every hop before finalUrl; redirectCount greater than 1 means a chain worth shortening.
• Run stopped early? RUN_STATS.stoppedByTimeBudget is true when the time guard was hit; raise the run timeout or lower maxUrlsChecked.
• Something isn't working? Open an issue on the Issues tab with the run ID.