Updated 2026-08-31
A phone reputation check is meant to answer whether a number belongs to a real, reachable person or to a disposable VoIP line bought for one signup. This guide covers the signals that genuinely answer that - line type, carrier, validity and messaging-app registration - explains why this API deliberately returns no reputation score, and is explicit that phone coverage here is in beta and narrower than email.
A phone lookup returns two things: number intelligence, and registration on the platforms that accept a phone number as input. Both are facts with a source, and neither is a rating.
{
"query": "+14155550123",
"queryType": "phone",
"matches": 2,
"results": [
{ "platform": "whatsapp", "registered": true, "method": "public_api", "checkedAt": "2026-08-31T09:12:04.1Z" },
{ "platform": "telegram", "registered": true, "method": "public_api", "checkedAt": "2026-08-31T09:12:04.3Z" }
],
"phoneDetails": {
"valid": true,
"carrier": "Verizon Wireless",
"lineType": "mobile",
"countryCode": "US"
},
"breaches": null,
"webMentions": null
}
// Note: no "riskScore", no "reputation". Those fields do not exist here.Read phoneDetails.lineType first. It carries more weight than the other three fields combined, for the reason below.
Numbers are not interchangeable. A postpaid mobile line is issued by a carrier that took identity and payment details for it, and it is expensive and slow to acquire in volume. A non-fixed-VoIP number can be provisioned in seconds, in bulk, from a browser, and discarded just as fast. When a fraud team asks whether a number is “good”, this is very nearly the whole question.
What line type does not tell you is whether the person is legitimate. Plenty of real people use VoIP numbers deliberately - remote workers, people who moved country, anyone who prefers not to hand out a personal mobile. It is a reason to want a second signal, not a reason to decline.
Only 5 platforms in the registry accept a phone number, and they are essentially the messaging apps. There is no version of this API where a phone number reaches the coverage an email address does, because the platforms themselves do not offer it - most have no way to ask “does an account exist for this number” at all.
So a phone registration result is a bonus signal rather than a foundation. A number registered on a messaging app is weak positive evidence that it belongs to a person; a number registered on none of them establishes very little, given how few were checked. When you hold both a number and an email address, run the address - see the email enrichment API. Phone lookups remain in beta, and the phone lookup guide sets out what that means.
Every input to a score is already in the response: validity, carrier, line type, country and registration. Turning them into a single number would add no information - it would only move the thresholds from your code into ours, where you cannot see them, cannot tune them for your population, and cannot explain them to a reviewer who asks why an application was declined.
Score it yourself, on the facts, and keep the facts beside the score. That is the same rule applied throughout: breaches and webMentions come back null when a source is not configured rather than empty, and a platform that could not be confirmed is omitted rather than reported as false. A number you cannot audit is worse than no number.