Pull emails, URLs, numbers and phone numbers out of text.
What it's for
Pull every email address, web link, mobile number or other item out of a long piece of text, as a clean list with each one written once.
Before you start
Have the text to copy from: an email thread, a chat export, a web page or your notes. It is read here in your browser and is not sent or kept anywhere.
Paste the text
Paste it into the Text box, replacing the sample. It can be as messy as you like — the tool finds the items wherever they are.
Choose what to find
In What to find, pick one kind: Email addresses, URLs (web links), Philippine mobile numbers, Numbers, Hashtags or Mentions (names that start with @). For two kinds, do it twice.
Sort the list, if you like
Tick Sort to put the list in alphabetical order. Leave it off to keep the order in which the items first appear in the text.
Extract and copy
Press Extract to bring the list into view — it also updates as you type. Press Copy at the top to take the whole list, one item per line.
Reading the result
If it goes wrong
This is the opposite end of the Text Redactor, and they share their patterns deliberately. One finds values in order to hide them; this finds them in order to collect them. If the two drift, this one sees something the redactor does not — the worst direction for that pair to disagree in.
It does two things a generic extractor does not. It normalises phone numbers: 0917 555 1234, +63 917 555 1234 and 639175551234 are one number written three ways.
And it does not swallow the sentence full stop. "See https://example.com/x." ends in a sentence, not in the link — and the stop taken along with it produces a URL that 404s.
0917 555 1234, +63 917 555 1234 and 639175551234 are one number written three ways. A list holding all three has not extracted anything useful, so they collapse to a single +63 form. You can switch normalisation off if you want them exactly as they appeared.
Because “See https://example.com/page.” ends in a sentence, and the stop is not part of the link — every naive extractor returns it, and the resulting URL 404s. Trailing sentence punctuation is trimmed. A closing bracket is kept when the URL opened one, because some links really do end in a bracket.
Because the use case is a list. Somebody pulling addresses out of a long thread wants each one once, not fifty times. The number of occurrences is shown beside each value, so collapsing them loses nothing, and you can turn deduplication off.
Because an @mention and the domain half of an email address start the same way. There is no reliable way to tell them apart from the text alone, so the page says so rather than silently guessing. Extract email addresses if that is what you are after.
It is deliberately broad — it matches years, quantities, parts of identifiers and anything else with digits in it. Narrowing it would silently drop the figure somebody was looking for, and a tool that quietly misses things is worse than one that returns a few extras.
No. Everything is found in your browser and nothing is transmitted or stored. That matters here because the text you paste to extract addresses from is usually somebody else’s correspondence.
A closing bracket is kept when the URL opened one. Real links do end in a bracket, and stripping it breaks them; keeping an unmatched one breaks them the other way.
Important
Runs entirely in your browser; nothing is transmitted. Patterns are deliberately broad — a few extra matches are better than silently missing the one you wanted. Philippine mobile numbers are normalised to +63 form unless you turn that off. Nothing you type here is sent to our servers — the calculation runs entirely in your browser.
