Text Extractor
Extract URLs, email addresses and numbers from text to turn unstructured copied content into useful lists.
Your result will appear here.
Text Extractor online
Pull every email address, URL and number out of a block of text — deduplicated, one per line — for harvesting the structured bits hiding in unstructured pastes.
How to use this tool
- Paste the source text — a document, a log, an exported page — into the input box.
- Click Run Tool. The text is scanned for URLs, email addresses and numbers.
- Copy the extracted list from the result panel.
Common uses
Harvesting contact addresses from a document or signature block, collecting links from a page export, pulling figures out of a report for a spreadsheet, and cleaning a data dump down to the values you actually need.
How it works
The extractor scans with three patterns in one pass: web addresses (http:// or https:// up to the next whitespace), email addresses (the standard local@domain shape with a real TLD), and numbers (integers and decimals). Matches are collected in the order they appear and deduplicated, so a signature repeated on every page of a document contributes its email once. The patterns are deliberately greedy-but-simple: a URL runs until whitespace, so trailing punctuation typed directly against it (a closing parenthesis, a full stop) is captured as part of the match.
Worked example
From the text Contact [email protected] or see https://example.com/help (updated 2026). We shipped 42 units at 9.5 kg. the extractor returns four lines: the email, the URL (with its closing parenthesis attached — trim it), and the numbers 2026, 42 and 9.5. One scan, three data types, no manual hunting — and a reminder to eyeball URL matches for the punctuation they picked up on the way.
Privacy and browser processing
Your text is scanned in your browser with JavaScript — it is never uploaded or stored.
Frequently asked questions
Why does my URL end with a ")" or "." ?
The URL pattern stops at whitespace, so punctuation typed flush against a link — very common in prose — joins the match. It is a one-character trim in the result. Recognising "the sentence ended" versus "the URL contains this" reliably needs more context than a pattern carries.
Why was a number inside an email or URL not listed separately?
Each stretch of text matches at most once, and the URL and email patterns are tried before the number pattern. The year inside a URL is part of that URL, not a separate number — the scan respects the larger structure first.
Does it extract phone numbers or dates?
Phone-shaped digit runs appear as their numeric fragments, not as assembled phone numbers, and dates surface as their component numbers. The extractor handles the three shapes it advertises; assembling semantic types (a date is three numbers plus intent) needs a parser for that type.
Are results in document order?
Yes — matches are listed in the order encountered, deduplicated by exact text. The same address written with different case counts as a different match, since the text differs.
Is my text uploaded?
No. Extraction runs in your browser; the source text never leaves the page — paste address-bearing documents with the usual care regardless.