API Documentation

Integrate AI audio, video & image detection into your applications with our powerful REST API. Get started in minutes with comprehensive documentation and code examples.

RESTful API
JSON Response
99.9% Uptime

Upload Media File

Detect AI-generated images & videos by uploading a file

POST https://api.deep3d.ai/v1/predict
Bearer ••••••••••••••••••••••••••••••••

Hover to reveal public token

curl -X POST https://api.deep3d.ai/v1/predict \
  -H "Authorization: Bearer ••••••••" \
  -F "file=@image.jpg"

# Or for video:
curl -X POST https://api.deep3d.ai/v1/predict \
  -H "Authorization: Bearer ••••••••" \
  -F "file=@video.mp4"

Detect from URL

Analyze images & videos directly from a URL

POST https://api.deep3d.ai/v1/predict-url
Bearer ••••••••••••••••••••••••••••••••

Hover to reveal public token

curl -X POST https://api.deep3d.ai/v1/predict-url \
  -H "Authorization: Bearer ••••••••" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com/image.jpg"}'

# Or for video URL:
curl -X POST https://api.deep3d.ai/v1/predict-url \
  -H "Authorization: Bearer ••••••••" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com/video.mp4"}'

Response Format

All API responses follow a consistent JSON structure

Image Response Example (/v1/predict)

JSON Response
200 OK
{
  "api_version": "v1",
  "results": {
    "prediction_info": {
      "predicted_label": "AI-Assisted",
      "flag": 2,
      "flag_label": "Portions of this content may have been produced using AI tools.",
      "ai_score": 42.685221803064145,
      "processing_time_ms": 1159.95
    },
    "meta_info": {
      "filename": "473588175_1141178511346166_7054045206363295695_n.jpg",
      "original_format": "JPEG",
      "size": [960, 960]
    }
  },
  "status": "success",
  "status_code": 200
}

Video Response Example (/v1/predict)

JSON Response
200 OK
{
  "api_version": "v1",
  "results": {
    "prediction_info": {
      "predicted_label": "Genuine",
      "flag": 0,
      "flag_label": "All content has been produced manually and may include edits.",
      "ai_score": 30.765525375609275,
      "processing_time": "52.28s"
    },
    "meta_info": {
      "source": "Device",
      "original_format": "VIDEO",
      "file_size_mb": 2.59,
      "size": [640, 360]
    }
  },
  "status": "success",
  "status_code": 200
}

Image URL Response Example (/v1/predict-url)

JSON Response
200 OK
{
  "api_version": "v1",
  "results": {
    "prediction_info": {
      "predicted_label": "AI-Generated",
      "flag": 1,
      "flag_label": "Portions of this content may have been produced with extensive AI editing.",
      "ai_score": 78.55305473009746,
      "processing_time_ms": 1153.01
    },
    "meta_info": {
      "filename": "NEW-WPD-BLOG-Top-Image-12.png?v=1695912820",
      "original_format": "PNG",
      "size": [1050, 600]
    }
  },
  "status": "success",
  "status_code": 200
}

Response Fields

predicted_label: Display label from backend ("Genuine", "AI-Generated", or "AI-Assisted")
flag: Numeric classification value
  • • 0 = Genuine/Real content
  • • 1 = AI-Generated content
  • • 2 = AI-Assisted content
flag_label: Human-readable message describing the detection result
ai_score: AI probability score (0-100, where higher values indicate more AI involvement)
processing_time_ms: Processing time in milliseconds (for images)
processing_time: Processing time as string, e.g., "52.28s" (for videos)

Image Formats

  • PNG, JPG, JPEG, WEBP
  • Max file size: 10MB

Video Formats

  • MP4, MOV, AVI, WEBM, MKV
  • Max file size: 30MB
  • Max duration: 2 minutes

API Features

Everything you need to integrate AI detection into your workflow

Fast Response

Average processing time under 5 seconds

Secure & Private

Images are not stored after processing

High Accuracy

99%+ accuracy on diverse image types

Multiple Formats

Supports images (JPEG, PNG, WEBP) and videos (MP4, MOV, AVI, WEBM, MKV)

99.9% Uptime

Enterprise-grade infrastructure

Easy Integration

Simple REST API with clear documentation