Protection

How sensitive you have asked us to be. Three presets, a custom setting, a per-number override, and a list of numbers that always get through.

Levels

A level sets how many calls from one number, or one run of numbers, are ordinary before they are not, and how long a call may run before you hear about it. The account has one level; any number can override it.

Level	Feels like
`relaxed`	A busy line where the same person calling back a few times is normal. Slower to react, fewer alerts.
`standard`	The default. Tuned on real floods against real agents.
`strict`	A number that should almost never hear from the same caller twice. Reacts fast, blocks for longer, and treats a number the network knows as worth a look on its own.
`custom`	Standard, with the thresholds you set.

The level is changed in the dashboard, under Settings → Scoring, by an admin. It is not writable through an API key, and neither is the trusted list: both change what we refuse on your behalf, and a leaked key should not be able to widen the door.

Read the level

GET/v1/protectionany key

The level, what it resolves to in numbers, the presets for comparison, and the trusted list.

Response

level"relaxed" | "standard" | "strict" | "custom": The account's level.
customobject: The overrides in force when the level is custom. Empty otherwise.
effectiveobject: The thresholds actually applied, whichever level set them. Fields below.
trustedstring[]: Numbers that score zero on every call. E.164.
presetsobject: The three presets, each in the same shape as effective.

Thresholds

exact_maxinteger: Calls from one number to one of yours within exact_window_min minutes before it is too many.
stem_maxinteger: Calls from one run of consecutive numbers within stem_window_min minutes before it is too many.
ceiling_mininteger: How long a call may run before a call.long_running alert. A number can set its own.
block_mininteger: How long the first block on a number lasts.
network_only_flagsboolean: Whether a number the network knows, with nothing else against it, is worth a review on its own.

RequestShell

BASE=https://api.murusai.com

curl "$BASE/v1/protection" \
  -H "Authorization: Bearer $MURUSAI_KEY"

RequestNode

const BASE = "https://api.murusai.com";
const KEY = process.env.MURUSAI_KEY;

const url = new URL("/v1/protection", BASE);
const res = await fetch(url, {
  headers: {
    Authorization: `Bearer ${KEY}`,
  },
});
const data = await res.json();

RequestPython

import os, requests

BASE = "https://api.murusai.com"
KEY = os.environ["MURUSAI_KEY"]

r = requests.get(
    f"{BASE}/v1/protection",
    headers={"Authorization": f"Bearer {KEY}"},
)
data = r.json()

200OKResponseJSON

{
  "level": "standard",
  "custom": {},
  "effective": {
    "exact_max": 3,
    "exact_window_min": 10,
    "stem_max": 5,
    "stem_window_min": 15,
    "ceiling_min": 20,
    "block_min": 15,
    "network_only_flags": false
  },
  "trusted": ["+14155550199"],
  "presets": {
    "relaxed": { … },
    "standard": { … },
    "strict": { … }
  }
}

Per number

A single number can sit at a different level from the account (a support line on relaxed, a sales line on strict) by setting protection when you add or update the number. It can also have its own ceiling_min. Nothing else is per-number.