Skip to content
Fotobots Partner Docs

Platform Overview

Welcome to the Fotobots Developer Portal. Fotobots is a high-throughput event photography and facial recognition engine engineered for mass participation events such as marathons, triathlons, and sporting festivals.

Core Architecture

Fotobots is designed around an asynchronous, event-driven architecture:

  1. Storage Tier (Cloudflare R2 / S3): High-durability object storage for high-resolution originals, web-optimized previews, and thumbnails.
  2. Indexing Queues (Cloudflare Queues): Asynchronous decoupled task queues that process incoming photos without blocking upload clients.
  3. AI Vision Containers: Deep-learning models that detect faces, compute 512-dimensional facial feature embeddings, and extract bib numbers via optical character recognition (OCR).
  4. Vector Database: High-speed nearest-neighbor similarity search matching runner selfies against millions of race photos in milliseconds.
┌──────────────┐     Presigned PUT     ┌──────────────┐
│ Partner App  ├──────────────────────►│ R2 / S3      │
└──────┬───────┘                       └──────┬───────┘
       │ POST /photos/photo/success           │
       ▼                                      ▼
┌──────────────┐     Enqueue Task      ┌──────────────┐
│ Fotobots API ├──────────────────────►│ CF Queue     │
└──────────────┘                       └──────┬───────┘
                                              ▼
                                       ┌──────────────┐
                                       │ AI Container │
                                       │ (InsightFace)│
                                       └──────────────┘

What You Need to Get Started

Before writing any integration code, ensure you have:

  1. A Fotobots Partner Account on the Web Admin portal (Sandbox Admin or Production Admin).
  2. An Organization API Key (e.g. fb_test_... or fb_live_...).
  3. An active Album ID representing your race or event.

👉 Next: Obtain and Configure your Authentication API Key →