Text Extractor

An all-in-one text extractor for contact details, dates and more

This text extractor scans any block of text and pulls out the structured details hidden in it — email addresses, web links, phone numbers, dates, times, IP addresses, hashtags, @mentions, Indian PIN codes, US ZIP codes and UK postcodes, plain numbers, or anything matching a regular expression you write. Tick the kinds you want and the results appear in one table, grouped by type, with the line each item came from.

Each value is also given a normalised form you can sort and match: phone numbers in international E.164 format (+919876543210), dates in ISO 8601 (2026-10-14), times on the 24-hour clock, IPv6 addresses in their standard compressed form, and links with https:// added where it was missing. Copy everything grouped by type, or download CSV or JSON for a spreadsheet or a script.

It runs entirely in your browser, which matters for the email threads, CRM exports and server logs people usually paste into a tool like this.

How to use the Text Extractor

1. Paste text into “Text to search”, or press “Open a file” to load a .txt, .csv, .log, .html or .eml file.
2. Under “What to extract”, tick the kinds of detail you need. “Select all” ticks every built-in type.
3. Set how to read ambiguous values: the country for phone numbers written without +91, +1 or +44, and whether 03/04/2026 is 3 April or March 4.
4. For anything else, tick “Custom pattern” and type a regular expression, with flags such as i for case-insensitive.
5. Tap a type above the table to show only those results, then copy them or download a CSV or JSON file.

Text Extractor features

What each type finds

How ambiguous text is resolved

Text is read in a fixed order — links, emails, IP addresses, dates, times, phone numbers, postal codes, mentions, hashtags, numbers — and a stretch of text taken by one type is not offered to the next. That is why “14/10/2026” is one date rather than three numbers, and why the 2001:db8 in an IPv6 address is not read as a time.

Numbers without a country code depend on your setting: with India selected, 020 7946 0958 is read as a Pune landline (+91 20…); with the UK selected it becomes a London number (+44 20…). A grouping such as 98765 43210 is always treated as an Indian mobile, and (415) 555-0132 as North American. Six-digit PIN codes and five-digit ZIP codes are only taken when a label, a state code or a place name gives them away, unless you allow any 5- or 6-digit number.

Writing a custom pattern

• Patterns use JavaScript regular-expression syntax: \b[A-Z]{3}-\d{4}\b finds ticket references like ABC-1234, and INV-\d+ finds invoice numbers.
• Add the flag i to ignore letter case, m so ^ and $ match at every line, s so a dot also matches line breaks and u for full Unicode.
• Tick “Return only capture group 1” to keep just the bracketed part of each match, such as the prefix in ([A-Z]{3})-\d{4}.
• Test and debug a pattern step by step in the Regex Tester, or let the Regex Generator build one for common formats.