Text Extractor - Extract Emails, URLs & Numbers
Extract email addresses, URLs, phone numbers, and numbers from any text. Free pattern extractor - paste and copy matches instantly.
Three steps to get started
Paste your text
Paste any text containing emails, URLs, phone numbers, or numbers into the input area - documents, web pages, CSVs, or any raw text.
Select a pattern type
Choose from Email Addresses, URLs, Numbers, or Phone Numbers tabs. The count for each type is shown on the tab.
Copy results
Copy individual items with the per-row button, or use "Copy all" to get the full list as newline-separated text.
Extract structured data from unstructured text
Text extraction is the process of scanning unstructured prose for recognizable patterns and pulling out every match as a clean list. Paste a page of email correspondence, a scraped HTML dump, a CSV export, or a copied PDF, and this tool returns the email addresses, links, phone numbers, and numeric values it contains - deduplicated and ready to paste into a spreadsheet.
The reason this is useful is that most real-world data arrives almost structured. A contact list lives inside a Word document. A set of endpoints is buried in a support thread. A quarterly figure sits mid-paragraph. Finding those by eye means scrolling and copying one item at a time, and the error rate on that job climbs fast past a few dozen items.
What each pattern actually matches:
- Emails - a local part of letters, digits, and
. _ % + -, an@, a domain, and a TLD of at least two letters. Plus-addressing (name+tag@example.com) and subdomains are handled. - URLs - anything beginning
http://orhttps://, captured through the full path, query string, and fragment. - Phone numbers - 10-digit North American style, with optional
+1, parentheses around the area code, and spaces, dots, or dashes as separators. - Numbers - integers and decimals, including negatives and grouped values like
1,250.75.
Duplicates, false positives, and what to check
Emails, URLs, and phone numbers are deduplicated automatically, so a signature block repeated forty times in a thread yields one address. Numbers are not deduplicated, because in a financial or scientific context a repeated value is meaningful data rather than noise.
Two limitations are worth anticipating. Bare domains without a scheme - example.com on its own - are deliberately skipped, since matching them would also catch ordinary sentences ending in .com or abbreviations with periods. And the numeric pattern is greedy by design: it will happily return version strings, dates, and ZIP codes alongside the figures you wanted, so skim the list before importing it. If your source is HTML, extracting URLs will also pick up asset paths and tracking links mixed in with the content links.
Common applications
- Marketing - extract email addresses from an exported contact list or newsletter archive
- Web scraping - pull all links from pasted HTML source code
- Data cleaning - isolate numeric values from mixed text for spreadsheet import
- Sales - collect phone numbers from a CRM export or business directory
- Research - extract citations (URLs) from a research paper or web article
- Development - find all API endpoints (URLs) mentioned in documentation
All extraction runs in your browser. No data is ever sent to a server.