FileGizmo Text tools
Extract URLs and links from text
Pull every link out of pasted text, with sentence punctuation trimmed off the end and duplicates removed. Works on prose and on bookmark exports.
Complete
Your result is ready
Simple by design
Extract links in three steps
Getting the links out of a block of text is easy to do badly, and the way it goes wrong is always the same: the link comes back with a full stop stuck to the end of it.
That happens because a full stop is a perfectly legal character in a URL, so any pattern that stops at whitespace will happily swallow the punctuation that ended the sentence. The result looks right in a list and fails when clicked. FileGizmo trims trailing sentence punctuation from every match, so a link at the end of a sentence, inside brackets, or wrapped in quotation marks comes back as the link and nothing else.
Detection covers anything beginning with http or https, and anything beginning with www. as well. Stopping there is a deliberate limit. In running prose, a pattern loose enough to catch example.com without a scheme is also loose enough to catch the end of a sentence that has no space after the full stop, along with version numbers, filenames and abbreviations. Being quietly wrong about which of those is a link is worse than missing a bare domain that the person can see for themselves.
The tool never fetches anything. It does not check whether a link resolves, follow a redirect, or ask a service to expand a shortened URL. That is not a limitation being apologised for. This page is served with a policy that permits no outbound connections at all, so the promise that your text is not being sent anywhere is enforced by the browser rather than asserted by us, and it is visible in any network panel.
Duplicates are removed by default. Matching is case-insensitive, because a host name is not case-sensitive, and the copy kept is the first one in the text rather than the last. Sorting is available for when the list is the deliverable, and off by default for when the order the links appeared in is part of what you are reading.
The pattern is a single negated character class with one quantifier, which means no nesting and therefore no catastrophic backtracking. A large paste costs time in proportion to its size.
The result says how many links were found and how many duplicates were dropped. Copy the list or download it as a text file.
- 1
Paste prose a bookmarks export or a copied page
- 2
Choose whether to remove duplicates and sort the list
- 3
Press Extract links and copy or download the result
Good to know
Frequently asked questions
Does a link at the end of a sentence keep the full stop?
No. Trailing sentence punctuation is trimmed, so a link followed by a full stop, comma, bracket or quotation mark comes back without it. The full stop belongs to the sentence.
Are links without a scheme found?
Yes, if they begin with www. Anything else is too ambiguous to detect reliably in prose without also matching ordinary words that contain a dot.
Does it follow or check the links?
No. Nothing is fetched. The tool reads the text you paste and never makes a request, which is what keeps the page honest about not contacting anyone.
What happens to duplicates?
They are removed by default, matching case-insensitively, and the first spelling in the text is the one kept.
Learn more