URL Extractor

How to extract every link from a page of text

Paste it on the left. Every link appears on the right, deduplicated, with the hosts listed underneath.

Two copy buttons for the two formats people need next: one per line for a spreadsheet, comma separated for anything that takes a list.

Tracking parameters are stripped

Everything starting with utm_ goes, along with the click identifiers the ad platforms attach: gclid from Google, fbclid from Facebook, msclkid, igshid, and the rest.

Those parameters describe how you arrived, not where you are going. Keeping them in an extracted list is how a competitor's analytics ends up with your campaign names in it, and how a document quietly records which email someone clicked through from.

Ordinary query parameters stay. ?id=42 and ?page=3 are part of the address, and a tool that strips those has broken the link rather than cleaned it.

The same page linked twice becomes one entry

This falls out of stripping the tracking, and it is often the most useful thing here.

A newsletter links the same article from the headline and again from the footer, with different campaign tags on each. Raw, that is two links. Cleaned, it is one, and the count at the top tells you how many collapsed.

Brackets in links

A closing bracket is kept when the link contains a matching opener, and dropped when it was wrapping the link inside a sentence.

That is not a detail. Wikipedia puts brackets in article paths constantly, so Kerning_(typography) is a real URL, and an extractor that strips the closing bracket hands you a link that does not resolve, with nothing in the output to suggest it was ever different. Meanwhile (see https://example.com/page) has to lose its bracket or the link is equally broken. The two cases look identical to a naive rule.

What counts as a link

A scheme, or a leading www.. Nothing looser.

The temptation is to match anything with a dot in it, which finds the links without a scheme and also finds config.json, v2.1 and the end of every abbreviated sentence. Precision is worth more than the handful of bare domains it misses.

What else is in that text

The email extractor does the same job for addresses, including the ones written as name at company dot com to avoid exactly this kind of matching.

If you are merging link lists from several places, the duplicate line remover shows which ones appeared more than once across the merge.

Everything runs in your browser, and no link is ever fetched. They are read out of the text, not visited.