Broken Link Quick Check — 5 Pages, 50 Links, Same-Origin

Supply up to 3 seed URLs you own or are authorized to audit. Each produces one sequential, bounded audit row with observed anchor/source-page evidence and target checks. Only HTTP 404/410 actually observed is labelled broken. A clean bounded sample does not prove the entire website has no broken links.

{"urls":["https://www.w3.org/Style/Examples/007/center-example.en.html"]}

Coverage and output

Per seed: at most 5 candidate page fetches (up to 5 same-origin redirects per fetch), discovery depth 2, 50 distinct canonical link targets, 100 anchor evidence occurrences, 4 MiB per body (a larger complete-2xx child page still counts as answered by its status, but its links are not read: bytes truncation; an oversized seed or partial 206 response is incomplete and free), and a 60-second total deadline. HTML must identify text/html or application/xhtml+xml. Only static a[href] links are discovered; relative URLs respect the first HTML base URL. Fragments are ignored. Paths and queries remain distinct. Canonical repeated seeds are audited once, within the raw 3-input cap.

pages lists HTML read. Join evidence[].target to checks[].target for source-page/anchor evidence of broken targets. counts reports observed targets, broken, unknown, excluded, and skipped external/invalid links. truncation names intentional page/depth/target/evidence caps, the site's own crawl delay (crawl_delay), or a deadline. Evidence includes non-broken links so checks are auditable. A target GET already observed during discovery supplies its status without a second HEAD check.

No external-origin checks (including redirect hops), browsers, JavaScript rendering, sitemaps, discovery beyond supplied seeds, state, monitoring, assets or form submission. If the seed itself lives on another origin (python.org → www.python.org, HTTP → HTTPS), the audit moves there once, re-reading that origin's robots.txt. After that, redirects stay on that exact origin: a link that redirects to another origin is listed as excluded (redirects_off_origin, with the observed status and destination) and is not followed. Every outgoing hop validates all resolved IP addresses, pins the connection to a validated public IP and keeps TLS hostname verification enabled. The honest user agent is Broken-Link-Quick-Check/0.1.

Robots exclusions are checked before page/target requests and redirect hops; advertised crawl delay is respected. A 429 answer is read the same way, as the site asking for a slower pace: from then on requests are 3 seconds apart, and the refused URL is asked once more. When the delay leaves no time for the remaining targets within the 60 seconds, they are listed as excluded (crawl_delay) without requests and the audit is still a completed, disclosed sample. A robots.txt that answers 4xx (other than 429), or one with no rules at all (only a Sitemap line, say), means no rules, as RFC 9309 specifies; Allow/Disallow lines outside any user-agent group are not guessed at. A 429, 5xx, unreachable or malformed robots.txt never means permission to crawl. Excluded targets are listed without requests. A denied seed is free. Unknown checks (403/429/5xx, TLS/transport/body failures, unsafe or robots-disallowed redirects) are not broken. Incomplete audits preserve available evidence and are uncharged. An audit that checked no link at all is uncharged too: a page with no static a[href] same-origin links (usually a JavaScript-rendered site, or one whose links are all http:// while the seed is https://) answers nothing about broken links.

Billing

$0.01 per completed bounded audit, one site-links-audited event per answered seed, not per page, link or broken result. That is the Free-plan price. Paid Apify plans get a Store discount: $9.00 per 1,000 on Starter, $8.00 on Scale and $7.00 on Business and above. Intentional disclosed coverage caps and robots exclusions may still leave a completed bounded audit billable; upstream-incomplete audits are free, even with useful retained evidence.

reasonCode	Meaning	Charged?
audited	Bounded audit completed, possibly no broken links or intentional coverage truncation	yes
invalid_url	Invalid/unsupported seed	no
unsafe_target	Unsafe DNS/IP seed or robots destination	no
robots_unreadable	robots.txt was 429/5xx/unreachable/malformed; no crawl	no
robots_denied	Seed excluded by robots	no
not_html	Seed did not identify HTML	no
no_links_checked	Audit finished but checked no same-origin link (none found, or every one excluded)	no
upstream_incomplete	Crawl or target checks incomplete; available evidence retained	no

Access and sample attribution

Caller is responsible for website terms, authorized access and content reuse rights. Public HTTP availability and robots allowance are not blanket permission over third-party content. This is diagnostic link/status evidence, not a page-content redistribution service. RFC 9309 explicitly distinguishes robots from authorization.

The prefill uses a public CSS example and its linked public examples. Sample attribution: Copyright © 1994–2025 World Wide Web Consortium (W3C®); copyright policy and document licence apply. The licence permits copies/portions for any purpose with source/copyright/status notices; no endorsement is implied. Sample source URLs accompany each evidence row. This permission is limited to that sample, not arbitrary customer sites.