How to Generate Realistic Fake Phone Numbers for Test Data

Ten random digits is not a phone number

US phone numbers follow the North American Numbering Plan (NANP), and the format XXX-XXX-XXXX has real constraints baked into it:

• Area code (NPA) — the first digit must be 2–9. 0 and 1 are reserved as trunk prefixes.
• Exchange / central office code (NXX) — same rule. First digit must be 2–9.
• N11 codes are reserved — 211, 311, 411, 511, 611, 711, 811, 911 are service codes, not area codes or exchanges.
• Line numb er — the last four digits are genuinely unconstrained, 0000–9999.

So 073-118-4492 is ten digits and it is not a phone number. Any validator worth using — libphonenumber, most form libraries, the majority of hand-rolled regexes — will reject it. Naive random generation produces an invalid area code roughly 20% of the time, and an invalid exchange another 20% on top of that.

If you're seeding a few hundred rows, that's a meaningful chunk of your fixtures silently failing.

A generator that actually respects the rules (Our Tool)

The wider point about test data

Phone numbers are a small example of a pattern that shows up everywhere in fixtures:

• Credit card numbers need to pass the Luhn checksum, or your payment form rejects them
• UUIDs have version and variant bits in fixed positions
• ISBNs, IMEIs, VINs, and IBANs all carry check digits
• Even em ail addresses have rules that asdf@asdf quietly violates

Random bytes shaped like the right thing are not the right thing. Any field with a validator on it needs a generator that knows the same rules the validator does — otherwise you're just generating a slower way to fail.

About DevToolStack

The phone generator is one of 118+ free tools on DevToolStack — a collection I built for exactly this kind of small, recurring developer task that doesn't justify writing a script.

Everything runs client-side in your browser. No accounts, no uploads, no limits, and nothing you paste is transmitted to a server — which matters more than it sounds when the thing you're decoding is a JWT from production.

A few that get the most use:

Generators — random Base64 and hex strings, UUIDs, PINs, colors, dates, user agents
Data Encoding — Base64, URL, HTML entity encode/decode
Formatters — JSON, SQL, XML
Text Diff Checker — Compare two blocks of text and see line-by-line differences highlighted instantly

Browse all the tools → https://devtoolstack.io/