Extract Phone Numbers

The Extract Phone Numbers tool scans any text for phone number patterns and returns a deduplicated list of every number found. Phone number extraction comes up constantly in contact management, lead processing, log audits, document mining, and data cleaning. The challenge is that phone numbers appear in wildly different formats — +1 (555) 123-4567, 5551234567, 555.123.4567, 555-123-4567, 1-800-FLOWERS, +44 20 7123 4567 — and a good extractor catches all of them while avoiding false positives on things that look like phone numbers but are not (like timestamps, product SKUs, or random digit sequences).

This tool uses pattern matching tuned for real-world phone number formats with optional country-specific parsing. Default mode catches US/Canada (NANP) numbers and common international formats. Country-specific mode narrows to a single format (US, UK, Germany, France, India, etc.) for cleaner results when you know the origin of the data. Output is normalized to E.164 (+15551234567) by default, which is the international canonical format used by Twilio, Stripe, and most modern APIs. Deduplicated, validated, and exportable to CSV, JSON, or plain list. Runs entirely in your browser so contact lists and communications logs stay private.

Phone number extraction is surprisingly tricky because valid formats vary dramatically by country and convention.

North American Numbering Plan (NANP) used by US, Canada, and most Caribbean: +1 followed by 10 digits, often written as (XXX) XXX-XXXX or XXX-XXX-XXXX. Starting digit can be 2-9; specific area codes and prefixes are reserved.

International (E.164): + followed by country code (1-3 digits) and subscriber number, total 15 digits max. Example: +442071234567 (UK, +44 country code).

European formats: vary by country. UK writes +44 20 7123 4567 (area + subscriber). France writes +33 1 23 45 67 89 or 01 23 45 67 89 domestically. Germany uses +49 30 12345678 or 030 12345678.

Common delimiters between digit groups: space, dash, period, slash, parentheses around area code. The regex normalizes these out for comparison.

Vanity numbers: 1-800-FLOWERS uses letters that map to digits (2=ABC, 3=DEF, etc.). The tool optionally converts letters to digits during extraction.

Avoiding false positives:
- Product SKUs like \"123-45-6789\" look like SSNs not phones
- Dates like \"01-15-2024\" look like phone patterns
- Timestamps like \"10:30:45\" are not phones
- Fax numbers labeled \"fax: ...\" should be distinguished from voice

The tool requires digit count ≥ 7 (smallest valid local number) and ≤ 15 (E.164 max) to avoid matching short sequences. It also requires either + prefix, country code, or standard delimiters to suggest phone format.

Validation beyond regex: check area code against valid NANP area codes, country code against ITU country code list, subscriber number length against country-specific rules. Full validation is complex but covers 95%+ of real-world numbers.

Normalization: convert the extracted number to E.164 by stripping delimiters and adding country code. Use libphonenumber-style rules (infer country from prefix or context). The tool applies a practical subset of libphonenumber rules.

Performance: regex extraction on multi-megabyte text runs in milliseconds. Validation adds negligible overhead.

Compared to writing a phone regex in code, this tool handles the many edge cases (international formats, vanity numbers, false positives) without requiring you to assemble the pattern yourself. For automated pipelines, libphonenumber is the gold standard; for interactive extraction, this tool is much faster.

Compared to contact manager import features, this is the extraction step before import. Run it to clean and dedupe a phone list, then load into HubSpot, Salesforce, or whatever CRM you use.

Compared to Unix grep with a phone regex, this tool has validation and E.164 normalization built in. grep is fine for raw pattern matching; this tool adds the post-processing you almost always need.