Skip to content
Fotobots Partner Docs

High-Volume Photo Ingestion Pipeline

When ingesting 50,000 to 200,000 photos during a marathon, attempting to stream image bytes directly through an API worker causes server timeouts and bandwidth bottlenecks.

Fotobots employs an asynchronous presigned storage pipeline:

Step 1: Get Upload Ticket        Step 2: Binary PUT            Step 3: Confirm Ingestion
┌─────────────┐                ┌─────────────┐                ┌─────────────┐
│ Partner App ├───────────────►│ Cloud S3/R2 ├───────────────►│ Fotobots API │
│ POST /signed│                │ Direct PUT  │                │ POST /success
└─────────────┘                └─────────────┘                └─────────────┘

Step 1: Request Presigned Upload URL

Call POST /openapi/photos/photo/signed with your albumId and file details:

Terminal window
curl -X POST https://sandbox-api.fotobots.run/openapi/photos/photo/signed \
  -H "Content-Type: application/json" \
  -H "x-api-key: fb_test_YOUR_KEY" \
  -d '{
    "albumId": "d3b07384-d113-4a37-b4d6-f3ec0a4b7f8c",
    "filename": "runner_1001.jpg",
    "contentType": "image/jpeg"
  }'

Response

{
  "uploadUrl": "https://storage.fotobots.run/r2/upload?X-Amz-Signature=...",
  "photoId": "9fa8b123-e456-4789-a012-3456789abcde"
}

Step 2: Direct Binary PUT to Storage

Execute a direct HTTP PUT request with the raw image binary data directly to the uploadUrl:

Terminal window
curl -X PUT "https://storage.fotobots.run/r2/upload?X-Amz-Signature=..." \
  -H "Content-Type: image/jpeg" \
  --data-binary "@runner_1001.jpg"

Step 3: Confirm Ingestion & Trigger AI Indexing

After the storage upload returns 200 OK, notify Fotobots to trigger the asynchronous facial recognition and bib OCR worker queues:

Terminal window
curl -X POST https://sandbox-api.fotobots.run/openapi/photos/photo/success \
  -H "Content-Type: application/json" \
  -H "x-api-key: fb_test_YOUR_KEY" \
  -d '{
    "albumId": "d3b07384-d113-4a37-b4d6-f3ec0a4b7f8c",
    "photoId": "9fa8b123-e456-4789-a012-3456789abcde"
  }'

Response

{
  "success": true,
  "message": "Photo queued for facial recognition and thumbnail processing"
}