Free Online Broken Link Checker
Scan a webpage for broken links and HTTP errors.
| Check | Link text | Destination | Type |
|---|
No links match the current filter.
Check results
| URL | Status | Result | Details |
|---|
Up to twenty links can be checked per run. The server issues lightweight HEAD requests with a short timeout, so a few 403 responses are normal on sites that block bots. A failed check does not always mean a broken link; it can mean the server refused an automated request.
About the Broken Link Checker
Broken links are quiet failures. A visitor clicks an outdated URL, lands on a 404 page, and leaves without telling anyone. Over time, reorganizations, renamed pages, expired domains, and deleted resources accumulate into a maze of dead ends that damages trust and wastes crawl budget. This tool extracts the links from a single page and checks a sample of them, giving you a quick signal before a larger crawl or a manual review.
How the check works
- The server fetches the page you enter and parses every anchor element.
- Links are resolved against the page URL, deduplicated, and classified as internal or external.
- You select up to twenty destinations. The first twenty are preselected for convenience.
- The server sends HEAD requests in parallel and reports the status code, success flag, or connection error for each one.
- Results can be exported as CSV for a spreadsheet or a ticket.
A HEAD request asks for headers without the body, which is fast and polite. Some servers do not implement HEAD correctly, and a few explicitly reject it while serving the same URL normally to a GET request. Others return 403 to any request that lacks a browser-like user agent, or 429 when a burst of parallel checks arrives. Treat those results as uncertain rather than definitive. A 404 or 410, on the other hand, is a strong signal that the resource is gone. Timeouts and DNS failures usually indicate a dead domain or a server that cannot be reached from this network.
Only the links on one page are analyzed. That keeps each run fast, but it does not replace a site-wide crawl. A useful routine is to check your most visited templates: the home page, a popular article, a product page, and a documentation index. If those look healthy, move to a crawler that follows internal links across the whole site. For large sites, schedule checks and compare exports over time so that new failures stand out from known, accepted ones such as login redirects that always require a session.
The twenty-link limit protects shared server resources and keeps the response time predictable. If you need to check more, run the tool in batches or point a dedicated crawler at your sitemap. The CSV export keeps the URL, status, success flag, and error text, so you can sort by status and assign fixes. Private addresses are rejected by the endpoint, and only public HTTP or HTTPS URLs are requested.
Frequently asked questions
Why do some links return 403 when they open fine in my browser?
Many sites reject automated clients, rate-limit bursts, or require cookies and JavaScript. A 403 from this checker means the server refused the HEAD request, not necessarily that the page is broken for a real visitor.
Are redirects counted as broken?
The endpoint follows up to five redirects by default and marks a final 2xx or 3xx response as OK. A chain that exceeds the limit or ends in a timeout is reported as a failure.
Can I check more than twenty links at once?
No. Each run is limited to twenty URLs to keep the service responsive. Repeat the scan with a different selection, or use a desktop crawler for a full audit.