Phone Number Validator

Phone Number Validator: bulk check, format, line type and carrier, offline

Paste a list of numbers, or point it at a field in one of your own Apify datasets. Every distinct number comes back as a row: valid or not and why not, the E.164, international, national and tel: forms, the country, the calling code, and whether it is a mobile or a landline.

Nobody is called and nothing is texted. This is a check against the world's numbering plans, so it tells you whether a number can be right at all, not whether it is switched on today.

📞 What Phone Number Validator does

Write the numbers however you have them. +44 20 7946 0018, (201) 555-0123, 0033 6 12 34 56 78 and tel:+81-3-1234-5678 all parse. A leading 00 becomes +, a tel: prefix is dropped, and spaces, brackets and dashes do not matter.

A number without + or 00 has no country in it, so the run uses your Default country to read it. (201) 555-0123 is a valid United States number and is nonsense as a French one, which is the whole reason that field exists.

For the numbers that pass, you also get the region the range belongs to, which for a landline is usually a town or city, the carrier the range was first handed to, and the time zones it covers. Region and carrier names come back in any of 33 languages.

The datasetId field is the part that saves real work. Point it at the output of a business-listing or lead scraper, name the field that holds the number, and the two run back to back without you exporting anything.

📥 What you give it

{
  "phoneNumbers": ["+44 20 7946 0018", "(201) 555-0123", "06 12 34 56 78"],
  "defaultCountry": "US",
  "onlyValid": true,
  "numberTypes": ["mobile"],
  "maxNumbers": 5000
}

Anything a filter leaves out is not charged, and neither are blanks or repeats. So onlyValid: true on a list that is half rubbish costs you roughly half.

In the United States and Canada, pick landline or mobile too if you are filtering by line type. Those two countries share number ranges between mobiles and landlines, so most numbers there honestly cannot be told apart and come back under that label.

📤 What you get back

A real row, from run fV91ws4Xze6pNDOfZ:

{
  "input": "+44 20 7946 0018",
  "valid": true,
  "possible": true,
  "reason": null,
  "e164": "+442079460018",
  "international": "+44 20 7946 0018",
  "national": "020 7946 0018",
  "rfc3966": "tel:+442079460018",
  "extension": null,
  "countryCode": "GB",
  "country": "United Kingdom",
  "callingCode": "+44",
  "numberType": "landline",
  "smsCapable": false,
  "region": "London",
  "carrier": null,
  "timeZones": ["Europe/London"],
  "source": "list",
  "sourceIndex": 1
}

carrier is empty on that row because it is a landline: carrier data covers mobile ranges.

Formats, region, carrier and time zones are filled in only for numbers that pass. An invalid number gets its verdict, its reason and nothing invented around it.

🧾 Reading the output

One row per distinct number, and that is the whole dataset. No refusal rows and no blank lines.

A sample row carries _sample: true and turns up only when you run with nothing to check.

Everything else about the run is in a key-value record called RUN_REPORT:

A run whose filters match nothing finishes normally and says so in the status message. A run with nothing to check at all fails.

▶️ How to run it

1. Open Phone Number Validator and click Try for free. 2. Paste numbers into Phone numbers, or pick a dataset under Dataset and name the field. 3. Set Default country if your list has local numbers in it. 4. Add filters under Only valid numbers, Only these line types or Only these countries. 5. Click Start, then download the dataset as JSON, CSV or Excel, or read it from the Apify API.

💰 How much does it cost?

$0.26 per 1,000 numbers checked. The same rate on every Apify plan, with no volume tiers.

One charge per row delivered. Blanks, repeats of a number already in the run, and anything your filters remove are not charged, and neither is the sample row. Region, carrier and time zones cost nothing extra.

💡 What people use it for

• Cleaning a scraped lead list down to the numbers that can actually be dialled.
• Splitting a list into mobiles and landlines before a texting campaign.
• Normalising a CRM export to E.164 so two systems finally agree on a format.
• Working out which countries a list is really from, by counting countryCode.
• Cutting a scraper's output straight into a clean sheet, using datasetId instead of a CSV.

🚧 What it does not do

• It does not call or text anything. No answer test, no ring test, no HLR lookup. Nothing tells

you whether a number is live, disconnected or answered today.

• No names, no owners, no WhatsApp check.
• carrier is the carrier the range was first given to. A ported number shows the old one. Take

it as a hint about the range, not as fact about the handset.

• United States and Canada numbers come back as "landline or mobile" most of the time, because

the ranges are genuinely shared. smsCapable is null for those.

• The numbering data has a date. A range that opened after the last update reads as not in use.
• A number not tied to one place lists every time zone of its country. A US toll-free number

names 42 of them.

• rfc3966 has no separators, so tel:+442079460018. That is how the standard form is written.
• 500,000 numbers in a run, and a large run wants the memory to match.

🧭 Which list-cleaning tool do you need?

❓ Questions people ask

Does it ring the number? No. Everything is worked out from the numbering plans, so nobody's phone lights up and nothing leaves the run.

What is the difference between valid and possible? possible says the length could be right. valid says the whole number matches a range that country actually uses. Filter on valid.

Why did a number I know is fine come back invalid? Nearly always a missing country. Set Default country to match where the local numbers were written, or put them in + form.

Can I check the numbers from another actor's run? Yes. Pick that run's dataset under Dataset and set the field name. Nothing needs exporting.

Is the carrier current? No. It is the carrier the range was originally allocated to. Porting is invisible to an offline check.

Do I pay for numbers my filters removed? No. Only rows that reach the dataset are charged.

🆘 If something breaks

Open the Issues tab on the actor page. Send the run ID and a number that came back wrong. The RUN_REPORT record in the run's key-value store carries the counts, the filters and the stop reason, which usually explains a short result on its own.