Extract emails, links, numbers and more from text

What counts as a match

Each type has a precise rule, so you know what to expect:

• Emails: a name, an @ sign and a domain with at least one dot and a final part of two or more letters. Punctuation around the address, such as angle brackets, quotes and a full stop at the end of a sentence, is not part of it.
• Links: anything that starts with http://, https://, ftp:// or www. and runs until a space, quote or angle bracket. A closing bracket or full stop at the end is dropped, but brackets inside the link are kept.
• Numbers: digits with an optional sign, thousands separators and decimals. 1,234.50, -7 and 3.14 are each one number.
• Hashtags and mentions: a # or @ followed by letters, digits or underscores, in any language. A hashtag needs at least one letter, so #2024 is skipped.
• Phone numbers: digit groups with the separators people usually type, covered below.

How phone detection works, and where it fails

There is no single phone number format, so this is a best guess. The tool looks for a run of digits with spaces, hyphens, dots, slashes or brackets between them, and accepts it if it has 7 to 15 digits, which is the range that covers national numbers up to the international maximum of 15. A number with a + or 00 at the start is accepted as it is. A plain run of digits needs 10 or 11 digits.

It rejects dates such as 2024-05-12, IP addresses such as 192.168.1.1, year ranges, ZIP+4 codes, thousands-grouped numbers and decimals. It will still miss numbers written in unusual layouts, and it will sometimes accept a long ID or a reference number. Treat the list as a starting point and check it.

Duplicates, case and order

Remove duplicates keeps the first copy of each item and drops later ones. Ignore case treats [email protected] and [email protected] as the same, which is how mail systems treat them in practice. It is on for everything except links, because the path of a link can be case sensitive.

The summary is worked out on every match, before duplicates are removed, so the counts show how often each domain, hashtag or mention appeared in your text.