Deepfake Detection

Documents & Biometrics

Flag synthetic or manipulated faces in selfies and video frames.

Analyse a selfie or short video for generative artefacts, face swaps and replay attacks and return a risk score, so liveness checks are not fooled by AI-generated media.

Endpoint

POST https://app.emver.io/api/sandbox/v1/deepfake-detection

Authorization: Bearer YOUR_API_KEY

Content-Type: application/json

Accept: application/json

Request parameters

Field	Type	Description
`media` required	file	Selfie image or video
`threshold` optional	number	Risk threshold (0–100)

Example request

cURL

curl -X POST https://app.emver.io/api/sandbox/v1/deepfake-detection \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json" \
  -d '{
    "media": "<base64>",
    "threshold": 0
}'

import requests

response = requests.post(
    "https://app.emver.io/api/sandbox/v1/deepfake-detection",
    headers={
        "Authorization": "Bearer YOUR_API_KEY",
        "Accept": "application/json",
    },
    json={
        "media": "<base64>",
        "threshold": 0
    },
    timeout=60,
)
response.raise_for_status()
data = response.json()

const response = await fetch("https://app.emver.io/api/sandbox/v1/deepfake-detection", {
  method: "POST",
  headers: {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
    "Accept": "application/json",
  },
  body: JSON.stringify({
      "media": "<base64>",
      "threshold": 0
  }),
});

const data = await response.json();

$response = Http::withToken('YOUR_API_KEY')
    ->acceptJson()
    ->post('https://app.emver.io/api/sandbox/v1/deepfake-detection', {
        "media": "<base64>",
        "threshold": 0
    });

$data = $response->json();

Replace YOUR_API_KEY and the placeholder values. Use the switch above to see production URLs.

Response

Structure · 200

{
  "status": INTEGER,
  "message": STRING,
  "data": {
    "is_deepfake": BOOLEAN,
    "risk_score": NUMBER,
    "signals": ARRAY
  }
}

Example · 200

{
  "status": 200,
  "message": "string",
  "data": {
    "is_deepfake": true,
    "risk_score": 0,
    "signals": "string"
  }
}

Error · 4xx / 5xx

{
  "status": INTEGER,
  "message": STRING
}

See Errors & status codes for what each status means and which calls are billed.

About this API

Send a selfie image or a short video clip. The API returns a deepfake probability, the attack types detected (face swap, synthetic generation, screen replay) and a pass or fail against your threshold.

Typical uses

Video KYC and selfie onboarding. Reject synthetic faces before an agent ever sees them.
Account recovery. Protect high-value resets from impersonation.

Questions

Does it replace Liveness Detection?: No. Liveness proves a real person is present; deepfake detection proves the face is not synthetic. Run both for high-risk journeys.