Phone number validator

It is worth being blunt about this, because most tools with this name are not.

This validator performs a syntactic check against published national numbering plans. It confirms that the number begins with a real country code, that the national part is a plausible length for that country, and — where the country publishes distinct prefix ranges — that the leading digits fall in a mobile or a landline range. Those three checks catch the overwhelming majority of numbers that are wrong: typos, missing digits, a trunk zero left in, a number entered without its country code.

It cannot tell you whether the line is in service. Nothing that runs in a browser can. Establishing that a number is live requires either placing a call or querying the operator's Home Location Register through a paid lookup service, and any free tool that shows you a green tick labelled "active" is showing you a syntactic result with a misleading word attached to it.

It also cannot tell you who owns the number, whether it has been ported to a different operator, or whether it is a virtual number. Number portability in particular has made prefix-based operator identification unreliable in most developed markets: a number that begins with a range historically assigned to one operator may have moved years ago. Prefix ranges still reliably distinguish mobile from fixed line, because those ranges are not portable across the boundary.

E.164 is the ITU standard that defines what an international telephone number is: up to fifteen digits, of which one to three are the country code, with no other punctuation. Written down it carries a leading plus sign to mark it as international rather than local.

The practical argument for storing numbers in E.164 is that it is the only unambiguous form. 020 7946 0958 means a London number to a British reader and nothing at all to a system in another country. +442079460958 means the same number everywhere, is directly dialable from anywhere, and sorts and deduplicates correctly.

The common mistake in the other direction is storing E.164 and displaying it. A person reading their own number wants to see it in the national format they recognise, so the right architecture is to store one canonical form and format for display — never the reverse.

The fifteen-digit ceiling is a hard limit of the standard, and a number longer than that is invalid regardless of what a form accepted. The shortest valid numbers are considerably shorter: some Pacific island states issue five-digit national numbers, which is why a length check has to be per country rather than a global minimum.

Every country's numbering plan reserves blocks of leading digits for particular purposes. In the United Kingdom, 07 is mobile and 01 and 02 are geographic fixed lines. In India, mobiles begin with 6, 7, 8 or 9. In Nigeria, 070, 080, 081, 090 and 091 are mobile. In much of continental Europe, 6 or 7 after the country code indicates mobile.

This still works because the mobile/fixed boundary is not portable: a number can move between operators, but it cannot move from a mobile block to a geographic one. Operator-level identification from a prefix, by contrast, has been unreliable in most markets since portability was introduced.

Why it matters for cost: in a large majority of countries, calling a mobile costs more than calling a landline, sometimes several times more. Knowing which one you have before you dial is the difference between an expected bill and a surprise. In a minority of countries — the United Kingdom, Germany and Kenya among them — the relationship is reversed and the mobile is cheaper, which is why our rate pages publish both columns rather than a single "from" figure.

Countries with a single unified plan, or where the assignment is not published in a separable form, produce an honest "cannot determine" here rather than a guess. The North American Numbering Plan is the largest such case: in the United States and Canada an area code does not distinguish a mobile from a fixed line at all, and only a network lookup can.

If you are validating numbers in bulk — a customer list, an event registration export, a spreadsheet inherited from somebody else — the useful sequence is: strip everything that is not a digit or a leading plus, add the country code where the number is local to a known market, remove the trunk prefix if one was left in, then check the length.

Doing it in that order matters. Removing the trunk prefix before you know the country removes the wrong digit for the countries whose numbers legitimately begin with 0 after the country code. Checking length before adding the country code flags every valid local number as too short.

Records that fail after all four steps are genuinely broken and should be corrected at the source rather than guessed at. A number one digit short is not recoverable by adding a zero somewhere plausible; it is a data entry error that needs the original.

For anything larger than a few thousand records, a library that embeds the full numbering-plan metadata is the right tool rather than a web page. This validator uses the same published plan data those libraries do, for one number at a time.