Broken link checker

Every link on one page, fetched and answered: the ones that 404, the ones that redirect, the ones that time out, with the words each link carries so you can find it in the page. A broken link is a visitor lost and a crawler sent nowhere, and on a small site each one is easy to find.

By , founder of Porteur · Updated 14 September 2026 · Markdown

What it checks

The tool reads one page, takes up to a hundred and fifty links from it, internal and external, and requests each one eight at a time: a HEAD request first, a small GET when the server refuses HEAD. Each link gets a status, the number of hops it went through, the address it finally landed on, and the time it took. Broken links (an error, or no answer) come first, with the words the link carries on the page, so you can find the sentence; redirecting links come second, with where they land, because a link on your own page that redirects is a hop you can remove by pointing at the final address. The counts separate your own site from elsewhere, and the list copies as tab-separated text.

Why broken links matter more than their number

A link that answers 404 is a dead end for a person and for a crawler. For the person, the page promised something and gave an error; a share of them leave the site. For the crawler, it is a page discovered through your link that does not exist, which is a small signal that the site is not maintained. The links that hurt most are your own: an internal link to a page you moved or removed sends your own visitors, and your own standing, nowhere.

Redirected links are the quieter cost. Each one works, one hop later, and each hop is a wait; a page whose links all redirect is a page whose author never updated it after a move. Point the links at the final address and the hops disappear.

How to read the result

  • The verdict counts the broken links, on your site and elsewhere, and the redirecting ones. Broken on your own site is the first fix; broken elsewhere is a link to update or drop.
  • Broken lists each failing link with its status, the words it carries on the page, and where it points. No answer means the host did not respond in eight seconds or is not public.
  • Redirecting lists the links that work one hop later, with the address they land on and the number of hops. On your own pages, change the link to the final address.
  • 403 or 429 on an external link often means the site refuses requests from tools, not from people; open it in a browser before changing anything.
  • Show all lists every link with its answer and the time it took; Copy the list puts the problems on the clipboard, one a line, for a ticket or a spreadsheet.

What to fix, in order

  1. Your own links first

    Internal links marked broken are entirely in your hands: fix the address or restore the page.

  2. Then the redirects on your own links

    Replace each with the final address. A site that moved keeps working through redirects for years; the links still deserve updating.

  3. Then external links that are gone

    Find the page that replaced it, or link to an archived copy, or cut the sentence.

  4. Add the redirect the other way

    If the broken internal link is to a page you removed, other sites may link to it too. A 301 from the old path to the closest live page keeps those.

Limits

One page a run, a hundred and fifty links, sixty seconds. Links behind a sign-in, links in menus drawn by script after load, and links a server hides from bots are not read or answer differently. Ten runs an hour from one address, because each run is up to a hundred and fifty requests to other people's servers.

Questions

Check my site, free

This reads the links on one page. The free check reads the links to your site: the pages other sites still link to that answer nothing, and the sites that link to your rivals and not to you.

  • Free check, no card
  • Read-only, your own accounts
  • Readable by your agent

Read next