~/docs cat faceswap-detection.md

Deepfake / face-swap

POST /api/defence/faceswap/upload-detect — detect face swaps and deepfaked faces in images.

Request: files (binary, required) — JPEG, PNG, or WEBP; multiple files per request supported.

curl

curl -X POST "https://api.scam.ai/api/defence/faceswap/upload-detect" \
  -H "x-api-key: <YOUR_API_KEY>" \
  -F "files=@/path/to/image.jpg"

response fields

success        boolean  request completed without error
task_id        string   unique task identifier
verdict        string   "fake" | "real"
confidence     float    0.0–1.0 fakeness score (higher = more likely fake)
num_faces      integer  faces detected in the image
ml_response    object   per-face detail:
  .results[].faces[].blended_fakeness_score   per-face score 0.0–1.0
  .results[].faces[].coordinates              bounding box (x/y min/max, px)
  .results[].faces[].expert_predictions       raw per-expert model scores
file_info      object   original_filename, file_size, mime_type, uploaded_at
credits_used   integer  1 per image

If no faces are detected, faces comes back as an empty array. This powers KYC/IDV selfie checks and deepfake detection flows.

← Quickstart AI-image detection →