API Reference

Predict the outcome of a verification

At signup, score the user’s phone number or email address (target) as legitimate or suspicious. Scoring-only — does not update counters by itself. When using Feedback, call predict before verification.started on the same target (and correlation_id when used) so feedback can warm Watch auth-start counters. Use Events for product fraud labels; use Feedback only if you run your own phone verification funnel outside Prelude Verify.

POST

watch

predict

Predict outcome

curl --request POST \
  --url https://api.prelude.dev/v2/watch/predict \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "target": {
    "value": "<string>"
  },
  "signals": {
    "ip": "127.0.0.1",
    "device_id": "<string>",
    "device_model": "<string>",
    "os_version": "<string>",
    "app_version": "<string>",
    "user_agent": "<string>",
    "ja4_fingerprint": "<string>",
    "is_trusted_user": true
  },
  "dispatch_id": "<string>",
  "metadata": {
    "correlation_id": "<string>"
  }
}
'

import requests

url = "https://api.prelude.dev/v2/watch/predict"

payload = {
    "target": { "value": "<string>" },
    "signals": {
        "ip": "127.0.0.1",
        "device_id": "<string>",
        "device_model": "<string>",
        "os_version": "<string>",
        "app_version": "<string>",
        "user_agent": "<string>",
        "ja4_fingerprint": "<string>",
        "is_trusted_user": True
    },
    "dispatch_id": "<string>",
    "metadata": { "correlation_id": "<string>" }
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    target: {value: '<string>'},
    signals: {
      ip: '127.0.0.1',
      device_id: '<string>',
      device_model: '<string>',
      os_version: '<string>',
      app_version: '<string>',
      user_agent: '<string>',
      ja4_fingerprint: '<string>',
      is_trusted_user: true
    },
    dispatch_id: '<string>',
    metadata: {correlation_id: '<string>'}
  })
};

fetch('https://api.prelude.dev/v2/watch/predict', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.prelude.dev/v2/watch/predict",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'target' => [
        'value' => '<string>'
    ],
    'signals' => [
        'ip' => '127.0.0.1',
        'device_id' => '<string>',
        'device_model' => '<string>',
        'os_version' => '<string>',
        'app_version' => '<string>',
        'user_agent' => '<string>',
        'ja4_fingerprint' => '<string>',
        'is_trusted_user' => true
    ],
    'dispatch_id' => '<string>',
    'metadata' => [
        'correlation_id' => '<string>'
    ]
  ]),
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>",
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.prelude.dev/v2/watch/predict"

	payload := strings.NewReader("{\n  \"target\": {\n    \"value\": \"<string>\"\n  },\n  \"signals\": {\n    \"ip\": \"127.0.0.1\",\n    \"device_id\": \"<string>\",\n    \"device_model\": \"<string>\",\n    \"os_version\": \"<string>\",\n    \"app_version\": \"<string>\",\n    \"user_agent\": \"<string>\",\n    \"ja4_fingerprint\": \"<string>\",\n    \"is_trusted_user\": true\n  },\n  \"dispatch_id\": \"<string>\",\n  \"metadata\": {\n    \"correlation_id\": \"<string>\"\n  }\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.prelude.dev/v2/watch/predict")
  .header("Authorization", "Bearer <token>")
  .header("Content-Type", "application/json")
  .body("{\n  \"target\": {\n    \"value\": \"<string>\"\n  },\n  \"signals\": {\n    \"ip\": \"127.0.0.1\",\n    \"device_id\": \"<string>\",\n    \"device_model\": \"<string>\",\n    \"os_version\": \"<string>\",\n    \"app_version\": \"<string>\",\n    \"user_agent\": \"<string>\",\n    \"ja4_fingerprint\": \"<string>\",\n    \"is_trusted_user\": true\n  },\n  \"dispatch_id\": \"<string>\",\n  \"metadata\": {\n    \"correlation_id\": \"<string>\"\n  }\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.prelude.dev/v2/watch/predict")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"target\": {\n    \"value\": \"<string>\"\n  },\n  \"signals\": {\n    \"ip\": \"127.0.0.1\",\n    \"device_id\": \"<string>\",\n    \"device_model\": \"<string>\",\n    \"os_version\": \"<string>\",\n    \"app_version\": \"<string>\",\n    \"user_agent\": \"<string>\",\n    \"ja4_fingerprint\": \"<string>\",\n    \"is_trusted_user\": true\n  },\n  \"dispatch_id\": \"<string>\",\n  \"metadata\": {\n    \"correlation_id\": \"<string>\"\n  }\n}"

response = http.request(request)
puts response.read_body

{
  "id": "<string>",
  "request_id": "<string>",
  "risk_factors": []
}

{
  "code": "<string>",
  "message": "<string>",
  "type": "<string>",
  "request_id": "<string>"
}

Call Predict at signup for a real-time legitimate or suspicious outcome on the identifier the user provides. Set target.type to phone_number or email_address with the corresponding value. Predict is scoring-only. It evaluates forwarded signals (IP, device, JA4, and more) and does not imply that you sent a message, started a verification step, or triggered any other side effect. It does not update long-lived abuse counters by itself. If you use Feedback for your own verification funnel, call Predict before verification.started on the same target and reuse metadata.correlation_id on both requests. Prelude links the feedback to the stored prediction so auth-start counters in the Watch pool receive those signals. Feedback alone without a prior predict only updates attempt-rate counters.

If you need…	Use
Product-specific fraud labels (`account.banned`, …)	Events
Counters for your own verification funnel (not Prelude Verify)	Feedback — after Predict on the same target

You can use Predict without Events or Feedback, or combine all three. See Introduction — How the pieces fit together.

Authorizations

Authorization

string

header

required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

target

object

required

The signup identifier to score — a phone number or email address.

Show child attributes

signals

object

The signals used for anti-fraud. For more details, refer to Signals.

Show child attributes

dispatch_id

string

The identifier of the dispatch that came from the front-end SDK.

Required string length: 36

Example:

"123e4567-e89b-12d3-a456-426614174000"

metadata

object

The metadata for this prediction.

Show child attributes

Response

string

required

The prediction identifier.

Example:

"prd_01jc0t6fwwfgfsq1md24mhyztj"

prediction

enum<string>

required

The prediction outcome.

Available options:

legitimate,

suspicious

request_id

string

required

A string that identifies this specific request. Report it back to us to help us diagnose your issues.

Example:

"3d19215e-2991-4a05-a41a-527314e6ff6a"

risk_factors

enum<string>[]

The risk factors that contributed to the suspicious prediction. Only present when prediction is "suspicious" and the anti-fraud system detected specific risk signals.

behavioral_pattern - The phone number past behavior during verification flows exhibits suspicious patterns.
device_attribute - The device exhibits characteristics associated with suspicious activity patterns.
fraud_database - The phone number has been flagged as suspicious in one or more of our fraud databases.
location_discrepancy - The phone number prefix and IP address discrepancy indicates potential fraud.
network_fingerprint - The network connection exhibits characteristics associated with suspicious activity patterns.
poor_conversion_history - The phone number has a history of poorly converting to a verified phone number.
prefix_concentration - The phone number is part of a range known to be associated with suspicious activity patterns.
suspected_request_tampering - The SDK signature is invalid and the request is considered to be tampered with.
suspicious_ip_address - The IP address is deemed to be associated with suspicious activity patterns.
temporary_phone_number - The phone number is known to be a temporary or disposable number.

Available options:

behavioral_pattern,

device_attribute,

fraud_database,

location_discrepancy,

network_fingerprint,

poor_conversion_history,

prefix_concentration,

suspected_request_tampering,

suspicious_ip_address,

temporary_phone_number

Example:

["suspicious_ip_address", "fraud_database"]

Send feedbacksOptional. Report verification-funnel steps (verification.started, verification.completed) when you run phone verification outside Prelude Verify. Feeds Watch abuse-rate counters for your own flow. Call Predict on the same target before verification.started and reuse metadata.correlation_id so auth-start counters receive predict signals; without a linked predict, only attempt-rate counters update on started. Not required if you only use Events and/or Predict, or if Verify already handles verification for that traffic.

Predict outcome

curl --request POST \
  --url https://api.prelude.dev/v2/watch/predict \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "target": {
    "value": "<string>"
  },
  "signals": {
    "ip": "127.0.0.1",
    "device_id": "<string>",
    "device_model": "<string>",
    "os_version": "<string>",
    "app_version": "<string>",
    "user_agent": "<string>",
    "ja4_fingerprint": "<string>",
    "is_trusted_user": true
  },
  "dispatch_id": "<string>",
  "metadata": {
    "correlation_id": "<string>"
  }
}
'

import requests

url = "https://api.prelude.dev/v2/watch/predict"

payload = {
    "target": { "value": "<string>" },
    "signals": {
        "ip": "127.0.0.1",
        "device_id": "<string>",
        "device_model": "<string>",
        "os_version": "<string>",
        "app_version": "<string>",
        "user_agent": "<string>",
        "ja4_fingerprint": "<string>",
        "is_trusted_user": True
    },
    "dispatch_id": "<string>",
    "metadata": { "correlation_id": "<string>" }
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    target: {value: '<string>'},
    signals: {
      ip: '127.0.0.1',
      device_id: '<string>',
      device_model: '<string>',
      os_version: '<string>',
      app_version: '<string>',
      user_agent: '<string>',
      ja4_fingerprint: '<string>',
      is_trusted_user: true
    },
    dispatch_id: '<string>',
    metadata: {correlation_id: '<string>'}
  })
};

fetch('https://api.prelude.dev/v2/watch/predict', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.prelude.dev/v2/watch/predict",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'target' => [
        'value' => '<string>'
    ],
    'signals' => [
        'ip' => '127.0.0.1',
        'device_id' => '<string>',
        'device_model' => '<string>',
        'os_version' => '<string>',
        'app_version' => '<string>',
        'user_agent' => '<string>',
        'ja4_fingerprint' => '<string>',
        'is_trusted_user' => true
    ],
    'dispatch_id' => '<string>',
    'metadata' => [
        'correlation_id' => '<string>'
    ]
  ]),
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>",
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.prelude.dev/v2/watch/predict"

	payload := strings.NewReader("{\n  \"target\": {\n    \"value\": \"<string>\"\n  },\n  \"signals\": {\n    \"ip\": \"127.0.0.1\",\n    \"device_id\": \"<string>\",\n    \"device_model\": \"<string>\",\n    \"os_version\": \"<string>\",\n    \"app_version\": \"<string>\",\n    \"user_agent\": \"<string>\",\n    \"ja4_fingerprint\": \"<string>\",\n    \"is_trusted_user\": true\n  },\n  \"dispatch_id\": \"<string>\",\n  \"metadata\": {\n    \"correlation_id\": \"<string>\"\n  }\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.prelude.dev/v2/watch/predict")
  .header("Authorization", "Bearer <token>")
  .header("Content-Type", "application/json")
  .body("{\n  \"target\": {\n    \"value\": \"<string>\"\n  },\n  \"signals\": {\n    \"ip\": \"127.0.0.1\",\n    \"device_id\": \"<string>\",\n    \"device_model\": \"<string>\",\n    \"os_version\": \"<string>\",\n    \"app_version\": \"<string>\",\n    \"user_agent\": \"<string>\",\n    \"ja4_fingerprint\": \"<string>\",\n    \"is_trusted_user\": true\n  },\n  \"dispatch_id\": \"<string>\",\n  \"metadata\": {\n    \"correlation_id\": \"<string>\"\n  }\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.prelude.dev/v2/watch/predict")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"target\": {\n    \"value\": \"<string>\"\n  },\n  \"signals\": {\n    \"ip\": \"127.0.0.1\",\n    \"device_id\": \"<string>\",\n    \"device_model\": \"<string>\",\n    \"os_version\": \"<string>\",\n    \"app_version\": \"<string>\",\n    \"user_agent\": \"<string>\",\n    \"ja4_fingerprint\": \"<string>\",\n    \"is_trusted_user\": true\n  },\n  \"dispatch_id\": \"<string>\",\n  \"metadata\": {\n    \"correlation_id\": \"<string>\"\n  }\n}"

response = http.request(request)
puts response.read_body

{
  "id": "<string>",
  "request_id": "<string>",
  "risk_factors": []
}

{
  "code": "<string>",
  "message": "<string>",
  "type": "<string>",
  "request_id": "<string>"
}