Integrating the Verisat Inference API

1. Prerequisites

• Enterprise or On-Premise plan subscription
• Access to the Verisat inference API endpoint
• API key (provided as part of your plan)
• Image URLs or files to analyze
• Kubernetes cluster (1.24+) or managed Kubernetes service
• Helm 3.x for deployment management
• Cloud provider credentials (AWS/GCP/Azure)
• Monitoring stack (Prometheus, Grafana)
• CI/CD pipeline (GitHub Actions, GitLab CI, etc.)

2. API Endpoint

3. Example Request

You can submit an image for analysis using curl:

curl -X POST https://sandbox.verisat.ai/api/v1/predict \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "data": [
      {"path": "https://example.com/image.png", "meta": {"_type": "FileData"}},
      {"type_of_analysis": "grid"}
    ]
  }'

Or using Python:

import requests

url = "https://sandbox.verisat.ai/api/v1/predict"
payload = {
    "data": [
        {"path": "https://example.com/image.png", "meta": {"_type": "FileData"}},
        {"type_of_analysis": "grid"}
    ]
}
headers = {
    "Content-Type": "application/json",
    "Authorization": "Bearer YOUR_API_KEY"
}

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

4. Analysis Type Parameter

The type_of_analysis parameter lets you specify the analysis mode:

• grid: Deep image analysis by grid (recommended for detailed forensics)
• overall: Overall image analysis (faster, less granular)

Example with overall analysis:

curl -X POST https://sandbox.verisat.ai/api/v1/predict \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "data": [
      {"path": "https://example.com/image.png", "meta": {"_type": "FileData"}},
      {"type_of_analysis": "overall"}
    ]
  }'

5. Sample Response

{
  "result": "fake",
  "confidence": 0.97,
  "details": {
    "regions_flagged": [
      {"x": 120, "y": 80, "width": 40, "height": 40, "score": 0.92}
    ]
  }
}

The response includes the overall result, confidence score, and (for grid analysis) flagged regions.

6. Error Handling

If your request is invalid or the image cannot be processed, you may receive an error response:

{
  "error": "Invalid image URL or unsupported file format."
}

Check your input and try again. For persistent issues, contact support.

7. Support

For help or to request additional features, email contact@verisat.ai.