
Paste notes, email threads or page source, strip tracking tags, dedupe, list unique domains and export a CSV of every URL it finds.
After a week of research, your notes are a swamp: a pasted email thread, half a bibliography, three chat exports and a page of HTML source copied from a competitor's article. Somewhere inside are the 40 links you need for the citation list. URL Extractor pulls them out of any pasted text and gives you a clean, deduplicated list you can copy or download as a CSV.
You paste, it lists. There is no upload step, and the scanning happens in your browser, so a paste of private notes is not sent anywhere to be processed.
Suppose your notes contain these lines: a health overview at an NIH address ending in ?utm_source=newsletter&utm_medium=email, a Wikipedia link whose address ends in a bracket pair (Sleep_(disambiguation)) that sits inside parentheses in your sentence, a bare www.cdc.gov/sleep/index.html, the same slides link twice, a DOI link, a news link carrying fbclid=AbC123 and ref=home, a mailing-list address, the word "example.com" typed alone, and a relative path.
The tool lists six entries. The slides link is merged because duplicates are removed. The Wikipedia link keeps its own closing bracket, because the pair is balanced, while the bracket that belonged to your sentence is trimmed. The www. link is listed as typed, without a protocol added. The email address, the bare domain and the relative path are all left out, since none starts with http, https, ftp or www.
id=77 and ref=home, are kept. Links that become identical after cleaning are merged.The first row is url,domain. Each following row holds an entry and its domain in lower case, without protocol, path or www. That layout opens cleanly in any spreadsheet, so you can add columns for title, access date and status, then sort by domain to spot where all your sources came from. If you turn on unique domains first, both columns hold the host name.
Because the CSV follows whichever switches are on at the moment you click, decide the shape before downloading. Do one download with full addresses for the citation list and, if useful, a second with unique domains for a quick source-diversity check.
You can paste page source straight in. The pattern stops at spaces, angle brackets and quotes, so addresses inside href and src attributes are picked up, which is handy for collecting every outbound link on a page you own. Relative links in those attributes, such as /about, are not captured because they carry no protocol, and neither are mailto links.
Expect noise: image, stylesheet and script addresses appear alongside article links. Use Unique domains only to see what the page loads from, or paste just the article body when you want citations. To grab the email addresses from the same paste, use Email Extractor instead, since this tool deliberately ignores them.
It never opens the links. A dead link and a live one look identical in the list, so spot-check important ones in a browser before you cite them. It will not rescue a bare domain: "example.com" on its own cannot be told apart from an ordinary word. A link glued directly to the preceding word or to an at sign is skipped, and anything shorter than eight characters is dropped.
It also leaves fragments alone. Two links to the same page with different #section endings are listed as two entries, because everything after the hash is kept. That is usually what a researcher wants, since the sections point to different passages.
Once you have your list, Text Sorter can remove duplicate lines and sort them if you merge lists from several colleagues. Keep the original paste alongside the CSV as well, so you can read the sentence around each link when you write the citation.
If you are tidying tracking parameters before publishing a newsletter, run the paste with tracking removal on, then spot-check three entries. Parameters you want to keep, such as a video timestamp, survive because only known tracking names are removed. More text utilities are in the text tools category, part of the 300+ professional tools, premium applications and workflow products on the site.
Paste your messiest notes file into URL Extractor, turn on tracking removal, and compare the count with the number of sources you thought you had.
Would a tool that also checks whether each extracted link still loads be useful in your workflow? Tell us on the contact page.