Validate phone numbers in Python

The plans in this file

Add the countries whose numbers you accept; the file and its tests below are generated again for them, in your browser.

• United Kingdom +44×
• United States +1×
• France +33×

1. The plans

PLANS holds one entry per plan: its region (the ISO code), the country code, the trunk prefix dialled before a national number, the exit code dialled before an international one, and a pattern per line type. The patterns are the plans' own, anchored, in the order the file tries them: the special ranges first (premium rate, toll-free, shared cost, VoIP, personal, pager, UAN, voicemail), then fixed line, then mobile. A plan that names no single exit code gets an empty one, and calls out of it use the + form.

"""phone.py: validate, normalise and dial phone numbers with the national numbering plans.

Plans: GB +44, US +1, FR +33 (data as of 2026-10-25).
Other plans: https://phonelookup.dev/guides/validate-phone-numbers-python?plans=gb,us,fr
Each pattern matches a national significant number: digits only, no country code, no trunk prefix.
"""
from __future__ import annotations

import re
from dataclasses import dataclass


@dataclass(frozen=True)
class Plan:
    region: str
    cc: str
    trunk: str
    exit: str
    types: tuple[tuple[str, re.Pattern[str]], ...]


@dataclass(frozen=True)
class PhoneNumber:
    e164: str
    region: str
    type: str
    cc: str
    nsn: str


def _types(*pairs: tuple[str, str]) -> tuple[tuple[str, re.Pattern[str]], ...]:
    return tuple((name, re.compile(pattern, re.ASCII)) for name, pattern in pairs)


PLANS: tuple[Plan, ...] = (
    Plan("GB", "44", "0", "00", _types(
        ("premiumRate", r"(?:8(?:4[2-5]|7[0-3])|9(?:[01]\d|8[2-49]))\d{7}|845464\d"),
        ("tollFree", r"80[08]\d{7}|800\d{6}|8001111"),
        ("voip", r"56\d{8}"),
        ("personalNumber", r"70\d{8}"),
        ("pager", r"76(?:464|652)\d{5}|76(?:0[0-28]|2[356]|34|4[01347]|5[49]|6[0-369]|77|8[14]|9[139])\d{6}"),
        ("uan", r"(?:3[0347]|55)\d{8}"),
        ("fixedLine", r"(?:1(?:1(?:3(?:[0-58]\d\d|73[0-5])|4(?:(?:[0-5]\d|70)\d|69[7-9])|(?:(?:5[0-26-9]|[78][0-49])\d|6(?:[0-4]\d|5[01]))\d)|(?:2(?:(?:0[024-9]|2[3-9]|3[3-79]|4[1-689]|[58][02-9]|6[0-47-9]|7[013-9]|9\d)\d|1(?:[0-7]\d|8[0-3]))|(?:3(?:0\d|1[0-8]|[25][02-9]|3[02-579]|[468][0-46-9]|7[1-35-79]|9[2-578])|4(?:0[03-9]|[137]\d|[28][02-57-9]|4[02-69]|5[0-8]|[69][0-79])|5(?:0[1-35-9]|[16]\d|2[024-9]|3[015689]|4[02-9]|5[03-9]|7[0-35-9]|8[0-468]|9[0-57-9])|6(?:0[034689]|1\d|2[0-35689]|[38][013-9]|4[1-467]|5[0-69]|6[13-9]|7[0-8]|9[0-24578])|7(?:0[0246-9]|2\d|3[0236-8]|4[03-9]|5[0-46-9]|6[013-9]|7[0-35-9]|8[024-9]|9[02-9])|8(?:0[35-9]|2[1-57-9]|3[02-578]|4[0-578]|5[124-9]|6[2-69]|7\d|8[02-9]|9[02569])|9(?:0[02-589]|[18]\d|2[02-689]|3[1-57-9]|4[2-9]|5[0-579]|6[2-47-9]|7[0-24578]|9[2-57]))\d)\d)|2(?:0[013478]|3[0189]|4[017]|8[0-46-9]|9[0-2])\d{3})\d{4}|1(?:2(?:0(?:46[1-4]|87[2-9])|545[1-79]|76(?:2\d|3[1-8]|6[1-6])|9(?:7(?:2[0-4]|3[2-5])|8(?:2[2-8]|7[0-47-9]|8[3-5])))|3(?:6(?:38[2-5]|47[23])|8(?:47[04-9]|64[0157-9]))|4(?:044[1-7]|20(?:2[23]|8\d)|6(?:0(?:30|5[2-57]|6[1-8]|7[2-8])|140)|8(?:052|87[1-3]))|5(?:2(?:4(?:3[2-79]|6\d)|76\d)|6(?:26[06-9]|686))|6(?:06(?:4\d|7[4-79])|295[5-7]|35[34]\d|47(?:24|61)|59(?:5[08]|6[67]|74)|9(?:55[0-4]|77[23]))|7(?:26(?:6[13-9]|7[0-7])|(?:442|688)\d|50(?:2[0-3]|[3-68]2|76))|8(?:27[56]\d|37(?:5[2-5]|8[239])|843[2-58])|9(?:0(?:0(?:6[1-8]|85)|52\d)|3583|4(?:66[1-8]|9(?:2[01]|81))|63(?:23|3[1-4])|9561))\d{3}"),
        ("mobile", r"7(?:457[0-57-9]|700[01]|911[028])\d{5}|7(?:[1-3]\d\d|4(?:[0-46-9]\d|5[0-689])|5(?:0[0-8]|[13-9]\d|2[0-35-9])|7(?:0[1-9]|[1-7]\d|8[02-9]|9[0-689])|8(?:[014-9]\d|[23][0-8])|9(?:[024-9]\d|1[02-9]|3[0-689]))\d{6}"),
    )),
    Plan("US", "1", "1", "011", _types(
        ("premiumRate", r"900[2-9]\d{6}"),
        ("tollFree", r"8(?:00|33|44|55|66|77|88)[2-9]\d{6}"),
        ("personalNumber", r"52(?:3(?:[2-46-9][02-9]\d|5(?:[02-46-9]\d|5[0-46-9]))|4(?:[2-478][02-9]\d|5(?:[034]\d|2[024-9]|5[0-46-9])|6(?:0[1-9]|[2-9]\d)|9(?:[05-9]\d|2[0-5]|49)))\d{4}|52[34][2-9]1[02-9]\d{4}|5(?:00|2[125-9]|3[23]|44|66|77|88)[2-9]\d{6}"),
        ("fixedLine", r"983[2-57-9]\d{6}|(?:2(?:0[1-35-9]|1[02-9]|2[03-57-9]|3[1459]|4[08]|5[1-46]|6[0279]|7[02469]|8[13])|3(?:0[1-57-9]|1[02-9]|2[013-79]|3[0-24679]|4[167]|5[0-3]|6[01349]|8[056])|4(?:0[124-9]|1[02-579]|2[3-5]|3[0245]|4[023578]|58|6[349]|7[02589]|8[04])|5(?:0[1-57-9]|1[0235-8]|20|3[0149]|4[01]|5[179]|6[1-47]|7[0-5]|8[0256])|6(?:0[1-35-9]|1[024-9]|2[03689]|3[016]|4[0156]|5[01679]|6[0-279]|78|8[0-269])|7(?:0[1-46-8]|1[2-9]|2[04-8]|3[0-2478]|4[0378]|5[47]|6[02359]|7[0-59]|8[156])|8(?:0[1-68]|1[02-8]|2[0168]|3[0-2589]|4[03578]|5[046-9]|6[02-5]|7[028])|9(?:0[1346-9]|1[02-9]|2[0589]|3[0146-8]|4[01357-9]|5[12469]|7[0-3589]|8[04-69]))[2-9]\d{6}"),
        ("mobile", r"983[2-57-9]\d{6}|(?:2(?:0[1-35-9]|1[02-9]|2[03-57-9]|3[1459]|4[08]|5[1-46]|6[0279]|7[02469]|8[13])|3(?:0[1-57-9]|1[02-9]|2[013-79]|3[0-24679]|4[167]|5[0-3]|6[01349]|8[056])|4(?:0[124-9]|1[02-579]|2[3-5]|3[0245]|4[023578]|58|6[349]|7[02589]|8[04])|5(?:0[1-57-9]|1[0235-8]|20|3[0149]|4[01]|5[179]|6[1-47]|7[0-5]|8[0256])|6(?:0[1-35-9]|1[024-9]|2[03689]|3[016]|4[0156]|5[01679]|6[0-279]|78|8[0-269])|7(?:0[1-46-8]|1[2-9]|2[04-8]|3[0-2478]|4[0378]|5[47]|6[02359]|7[0-59]|8[156])|8(?:0[1-68]|1[02-8]|2[0168]|3[0-2589]|4[03578]|5[046-9]|6[02-5]|7[028])|9(?:0[1346-9]|1[02-9]|2[0589]|3[0146-8]|4[01357-9]|5[12469]|7[0-3589]|8[04-69]))[2-9]\d{6}"),
    )),
    Plan("FR", "33", "0", "00", _types(
        ("premiumRate", r"836(?:0[0-36-9]|[1-9]\d)\d{4}|8(?:1[2-9]|2[2-47-9]|3[0-57-9]|[569]\d|8[0-35-9])\d{6}"),
        ("tollFree", r"80[0-5]\d{6}"),
        ("sharedCost", r"8(?:1[01]|2[0156]|4[024]|84)\d{6}"),
        ("voip", r"9\d{8}"),
        ("uan", r"80[6-9]\d{6}"),
        ("fixedLine", r"(?:26[013-9]|59[1-35-9])\d{6}|(?:[13]\d|2[0-57-9]|4[1-9]|5[0-8])\d{7}"),
        ("mobile", r"(?:6(?:[0-24-8]\d|3[0-8]|9[589])|7[3-9]\d)\d{6}"),
    )),
)

2. Reading a number

parse keeps the digits and a leading plus. With the plus, or with the default region's exit code in front (00 from the United Kingdom, 011 from the United States), the first one to three digits are the country code: under ITU-T E.164 no code is the prefix of another, so at most one length names a plan, and the plans sharing a code (+1, +7, +44) are tried in turn. Without either, the number is national and is read in the default region's plan. The trunk prefix is tried stripped first, so “+44 (0)7400 123456” reads too. A number in both the fixed-line and the mobile ranges comes back as fixedLineOrMobile, which is every ordinary number in the +1 plans. Digits typed in other scripts are dropped with the punctuation; convert them first if your users type them. Anything no plan accepts gives None.

_BY_REGION = {plan.region: plan for plan in PLANS}


def parse(text: str, default_region: str | None = None) -> PhoneNumber | None:
    """Read a number the way people type it: "+44 7400 123456", "07400 123456", "00 44 7400 123456".

    A number without + or an exit code is read in the plan of default_region (an ISO code such
    as "GB"). Returns None when no plan in PLANS accepts it.
    """
    s = re.sub(r"[^0-9+]", "", text)
    digits = s.replace("+", "")
    if not 3 <= len(digits) <= 17:
        return None
    if s.startswith("+"):
        return _international(digits)
    home = _BY_REGION.get(default_region or "")
    if home is None:
        return None
    if home.exit and digits.startswith(home.exit):
        return _international(digits[len(home.exit):])
    return _match([home] + [p for p in PLANS if p.cc == home.cc and p is not home], digits)


def _international(digits: str) -> PhoneNumber | None:
    # Country codes are prefix-free (ITU-T E.164): the first one to three digits that name a plan are the code.
    for size in range(1, min(3, len(digits)) + 1):
        plans = [p for p in PLANS if p.cc == digits[:size]]
        if plans:
            return _match(plans, digits[size:])
    return None


def _match(plans: list[Plan], nsn: str) -> PhoneNumber | None:
    # The trunk prefix is tried stripped first: "+44 (0)7400 123456" is common.
    for plan in plans:
        tries = [nsn[len(plan.trunk):], nsn] if plan.trunk and nsn.startswith(plan.trunk) else [nsn]
        for national in tries:
            kind = _type_of(plan, national)
            if kind:
                return PhoneNumber(f"+{plan.cc}{national}", plan.region, kind, plan.cc, national)
    return None


def _type_of(plan: Plan, nsn: str) -> str | None:
    # The first matching type wins; a number in both the fixed-line and the mobile ranges is reported as such.
    hits = [name for name, pattern in plan.types if pattern.fullmatch(nsn)]
    if not hits:
        return None
    if hits[0] == "fixedLine" and "mobile" in hits:
        return "fixedLineOrMobile"
    return hits[0]

3. Writing and dialling it

Store the E.164 form, a plus and digits only: one string per number, and every other format follows from it.to_tel writes the tel: URI for links (RFC 3966). dial_from gives the digits a person dials from another country: the caller's exit code, the country code and the national number, or, between plans that share a country code, a national call with the trunk prefix. The dialling helper explains each case.

def to_tel(number: PhoneNumber) -> str:
    """The tel: URI (RFC 3966), for links."""
    return f"tel:{number.e164}"


def dial_from(number: PhoneNumber, from_region: str) -> str:
    """What to dial from a phone in from_region.

    Inside one country code (the UK and Jersey, the US and Canada) the call is dialled like a
    national one, with the trunk prefix; otherwise the exit code, the country code and the national
    number. The + form when from_region is not in PLANS or its plan names no exit code.
    """
    origin = _BY_REGION.get(from_region)
    if origin is None:
        return number.e164
    if origin.cc == number.cc:
        return origin.trunk + number.nsn
    return origin.exit + number.cc + number.nsn if origin.exit else number.e164

The tests

The cases are the plans' own example numbers, one per line type in international form; the first plan's mobile number again in national form, with “(0)”, and dialled with an exit code; then numbers no plan accepts, and three calls dialled. Their expected values are computed when the page is built, by the same logic in TypeScript. The generator behind this page is checked by running its output in unittest (pytest runs it too), for these plans and for all 245 plans at once.

python -m unittest

import unittest

from phone import dial_from, parse, to_tel


class PhoneTest(unittest.TestCase):
    def test_reads_each_line_type_of_each_plan(self):
        cases = [
            # typed, default region, E.164, line type
            ("+44 7400 123456", None, "+447400123456", "mobile"),
            ("+44 121 234 5678", None, "+441212345678", "fixedLine"),
            ("+44 800 123 4567", None, "+448001234567", "tollFree"),
            ("+44 56 1234 5678", None, "+445612345678", "voip"),
            ("+44 901 234 5678", None, "+449012345678", "premiumRate"),
            ("+44 70 1234 5678", None, "+447012345678", "personalNumber"),
            ("+44 76 4012 3456", None, "+447640123456", "pager"),
            ("+44 55 1234 5678", None, "+445512345678", "uan"),
            ("+1 201-555-0123", None, "+12015550123", "fixedLineOrMobile"),
            ("+1 800-234-5678", None, "+18002345678", "tollFree"),
            ("+1 900-234-5678", None, "+19002345678", "premiumRate"),
            ("+1 500-234-5678", None, "+15002345678", "personalNumber"),
            ("+33 6 12 34 56 78", None, "+33612345678", "mobile"),
            ("+33 1 23 45 67 89", None, "+33123456789", "fixedLine"),
            ("+33 801 23 45 67", None, "+33801234567", "tollFree"),
            ("+33 9 12 34 56 78", None, "+33912345678", "voip"),
            ("+33 884 01 23 45", None, "+33884012345", "sharedCost"),
            ("+33 891 12 34 56", None, "+33891123456", "premiumRate"),
            ("+33 806 12 34 56", None, "+33806123456", "uan"),
            ("07400 123456", "GB", "+447400123456", "mobile"),
            ("+44 (0)7400 123456", None, "+447400123456", "mobile"),
            ("00 1 201-555-0123", "GB", "+12015550123", "fixedLineOrMobile"),
        ]
        for text, region, e164, kind in cases:
            with self.subTest(text=text):
                number = parse(text, region)
                self.assertIsNotNone(number)
                self.assertEqual(number.e164, e164)
                self.assertEqual(number.type, kind)

    def test_rejects_what_no_plan_accepts(self):
        self.assertIsNone(parse("12", "GB"))
        self.assertIsNone(parse("+44 7400 12345"))
        self.assertIsNone(parse("+49 1512 3456789"))
        self.assertIsNone(parse("07400 123456"))

    def test_writes_and_dials(self):
        self.assertEqual(to_tel(parse("+44 7400 123456")), "tel:+447400123456")
        self.assertEqual(dial_from(parse("+1 201-555-0123"), "GB"), "0012015550123")
        self.assertEqual(dial_from(parse("+44 7400 123456"), "US"), "011447400123456")
        self.assertEqual(dial_from(parse("+44 7400 123456"), "GB"), "07400123456")


if __name__ == "__main__":
    unittest.main()

What is specific to Python

• fullmatch, not match: re.match anchors only at the start, so a pattern for eleven digits also accepts eleven digits followed by anything.
• The patterns compile with re.ASCII. In a str pattern, \d otherwise matches any Unicode decimal digit, Arabic-Indic and Devanagari included.
• $ matches before a trailing newline (re.match(r"^\d+$", "123\n") succeeds); fullmatch has no such gap.
• The patterns compile once, at import; PhoneNumber and Plan are frozen dataclasses, safe to share between threads.