Paste any text and get its web links, one per line. Nothing leaves your browser.
A link starts with http:// or https:// (any capitals) followed by a letter or digit, and runs until the next space, line break, angle bracket, quotation mark, apostrophe or backtick. So everything in the address is kept: the path, the query after the question mark and the part after the hash. Other schemes such as ftp:// or mailto: are not found, and a bare http:// with nothing after it is not a link. Letters from any alphabet are accepted, so https://münchen.de/straße is found whole.
Sentences put punctuation right after a link, so the tool removes trailing dots, commas, semicolons, colons, exclamation marks, question marks, apostrophes and asterisks. A closing parenthesis, square bracket or brace is removed only when it has no opener in the link. That is why (see https://example.com/a) gives https://example.com/a, while https://en.wikipedia.org/wiki/Tab_(x) keeps its final bracket. A link that really ends in a question mark or a dot cannot be told apart from a sentence ending, so those characters are always trimmed. Check such links by hand.
Include www. links without a scheme is on by default. It also finds addresses that start with www. and a letter or digit, such as www.example.org/news. They are listed as written, without http:// added, so paste them into a browser address bar or add the scheme yourself. A www. that sits inside a longer word, an email address or a path is ignored, and a www. inside an http link is part of that link and not listed twice. Unique only is on by default and compares links exactly, including capitals, since paths can be case sensitive: https://a.com/x and https://a.com/X are two links, and so are a link with and without a final slash. Sort A to Z orders the list ignoring case.
The search runs in this tab and nothing is uploaded or stored apart from your three choices. Text up to 1,000,000 characters is handled. The tool only reads the text; it does not visit the links, check that they work or expand shortened links. A text with no links shows a short note instead of an empty box.
No. It only finds text shaped like a link. It never visits the address, so a dead link or a typo such as htps:// versus https:// is listed or missed purely by its shape.
A final dot, comma, question mark or similar mark is treated as sentence punctuation and trimmed. A closing bracket is trimmed too unless the link opened one. Add the character back by hand in the rare case it was part of the address.
They are copied as they appear in the text. If the text said www.example.org, that is what you get. Turn the option off to list only links that carry http:// or https://.
Yes. Paste the source or the markdown. Quote marks and angle brackets end a link, so href values and links in angle brackets work. In [name](https://...) only the address is returned.
With Unique only on, two links are the same only if every character matches, including capitals and a final slash. Links that differ by a tracking tail such as ?utm=1 count as different.