menu_book

Developer documentation

Everything you need to integrate real‑time object detection — REST API, WebSocket streams, edge deployment, and client SDKs.

search ⌘K

bolt Quickstart — 5 minute integration

Get your first detection in under a minute using our REST API. All requests require an API key.

                curl -X POST
                https://api.espyeye.io/v1/detect \

                -H
                "Authorization: Bearer YOUR_API_KEY"
                \

                -F "image=@street_view.jpg"
              

Response includes labels, confidence, bounding boxes

vpn_key Authentication

All API requests require an API key passed via the Authorization header. You can generate keys from the dashboard.

Header: Authorization: Bearer ee_abc123...xyz

✨ API keys are environment‑aware: use test_ prefix for sandbox.

sensors WebSocket live stream

Establish a persistent connection for real‑time detection with sub‑100ms latency.

                const ws =
                new WebSocket("wss://api.espyeye.io/v1/stream?key=YOUR_API_KEY");

                ws.onmessage = (event) => {

                  const detections = JSON.parse(event.data);

                  console.log(detections);
                // [{label: "person", confidence: 0.98, bbox: [...]}]

                };

code Python SDK (pip)

Install our official Python library for easy integration.

pip install espyeye-sdk

                from espyeye
                import Client

                client = Client(api_key="YOUR_KEY")

                result = client.detect_image("path/to/frame.jpg")

                print(result.objects)

router Deploy at the edge

Run models directly on NVIDIA Jetson or any Linux edge device. Our edge runtime containers are optimized for ARM64 and CUDA.

                docker run -d --runtime=nvidia \

                -e API_KEY="YOUR_KEY" \

                -p 8080:8080 \

                espyeye/edge-runtime:latest

→ then stream RTSP or local camera directly to the container.

API endpoints overview

Endpoint	Method	Description
/v1/detect	POST	single image detection
/v1/detect/batch	POST	batch up to 32 images
/v1/stream	WebSocket	persistent video stream
/v1/models	GET	list available models
/v1/health	GET	service status

Still need help?

We're here — chat with our developer advocates.

chat Discord mail Support

Ready to build?

Get your API keys instantly — free tier includes 1,000 detections/month.