One socket gives you a live verdict on every call.verdict.score 0.93
verdict.level "scam"
verdict.category "Calls pretending to be government, businesses, or family and friends"
You decide what to do with it. The filter never touches the call.What you get#
| Field | Meaning |
|---|
score | 0 to 1. The probability that the caller is a scammer. |
level | clean, suspicious, likely_scam, or scam. |
category | One exact CallerAPI complaint subject, or none. |
tactics | The manipulation techniques the caller used. |
signals | Short pieces of evidence. Above a score of 0.6 they quote the caller. |
summary | Two sentences at most. |
intel | Callback numbers, names, payment rails, remote tools, and other facts the caller exposed. |
With report=true, a final level of scam files the caller's number and every callback number as a complaint. That block applies across CallerAPI. See Block a caller.What it does not do#
It does not answer the call and it does not speak. The honeypot does that. It does not prove fraud. A verdict is a probability. scam means a score of 0.85 or more.
The four steps#
Rules that apply everywhere#
Send phone numbers in E.164 format. A plus sign, the country code, then the number. +15559990000 is correct. (555) 999-0000 is not. A number without a leading + is never filed.Send the caller's audio only. Do not mix both parties into one channel. The verdict is about the caller.Act on level, not on score. The levels are stable. The score moves. Modified at 2026-09-18 16:56:37