Somebody needs a list of every link on a page, and the honest options are a crawler you have to install or an afternoon with Ctrl+F. Give it the address instead, or paste the source yourself.
The quickest is to type the address and let the page fetch it. That is the only part of this site that talks to a server, so it is worth knowing what happens: the address goes to us, we request the page, the markup comes back to you. Nothing is kept.
The second is to view the page source, select it all and paste it. Same result, nothing leaves your computer, and it works on pages behind a login that our server could never reach.
The third surprises people. Select just the part of the page you care about, copy, and paste that: your clipboard carries the markup behind a selection as well as the words. One article, one table, one sidebar, with its links and nothing else. Handy when a page has four hundred links and you want the eleven in the body copy.
Most links on a page are not written as full addresses. They are written as /pricing/ or ../contact.html, because the browser already knows which page it is on and can work out the rest. Take that markup somewhere else and the context is gone with it.
This is the one place the two ways in differ. Fetch a page and the address is known, so every relative link is resolved into a real one exactly as the browser would have done, and your own links are told apart from everybody else's. Paste the markup and there is nothing to resolve against, so the links come back precisely as they were written. Neither is wrong; they answer different questions.
This is the kind of job you can hand to a chatbot, and for a short page it will do fine. On a long one it will sometimes give you thirty-eight of the forty-one links and sound entirely confident about it, because summarising is what it is built to do and a list is the one place that instinct is wrong.
A parser has no instinct. It returns every anchor with an href, in document order, and the count at the bottom is the count. When the output is going into a redirect map or a broken link check, that difference is the only thing that matters.
Three ways. Type the address and let the page fetch it, which is quickest. View the page source, select all and paste it, which keeps everything on your own computer and works on pages behind a login. Or select just the part of the page you care about and copy that, because your clipboard carries the markup behind a selection as well as the words. The third way is how you get the eleven links in an article without the four hundred around it.
Only in one case, and only one thing. If you paste the markup, nothing is sent anywhere: the whole job happens in your browser and you can watch the Network tab stay empty. If you ask the page to fetch an address, that address is sent to this server, the server requests the page and hands the markup back to you. It is not logged and not stored. The reason it cannot be done in your browser is that a page on another domain is off limits to JavaScript unless that domain has invited us, and almost none have.
Because that is how they were written. A browser knows which page it is on, so a link can leave out everything the browser already knows. Fetch the page instead of pasting it and this is solved for you: the address is known, so every relative link is resolved into a full one, exactly as the browser would have done. Pasted markup carries no such address, so it is read exactly as written.
By hostname, when it knows one. Fetch a page and it does: anything on that host is same-site and the rest is off-site. Paste markup and there is no host to compare against, so it falls back on how the links are written, which is a decent approximation: a link written as /about/ was meant for the same site, and one written out in full was usually not.
No, and that is deliberate. It returns every anchor element with an href attribute, which is what a link is. A button wired up with JavaScript, a URL sitting in a data attribute or an address written in plain prose are all different things, and guessing at them would make the count untrustworthy. The number at the bottom is exact.
Yes. Choose CSV and you get a proper two-column file with the addresses and the link text, quoted correctly, which opens without an import wizard. The URL and link text setting gives the same two columns separated by a tab, which is what you want when pasting straight into an open sheet.
No. That would mean requesting every address on the list, which is a different job with a different cost: hundreds of requests to other people's servers, sent from ours, on your behalf. What you get here is an accurate list of what the page links to, which is the tedious half. Feeding that list to a link checker afterwards is the easy half.
Any address written as a mailto: link, yes, and they are included by default. Addresses written as plain text in the middle of a sentence are not links, so they are not part of this tool's job.