Introduction

Raeh turns raw sensor data from your wearable into live health insights (heart rate, SpO₂, respiratory rate) delivered as JSON over a WebSocket.

Your device pushes PPG (and optionally accelerometer) samples to /stream/ingest. We run the signal processing pipeline (denoising, motion-artifact rejection, peak detection, HRV) and publish insights on /stream/subscribe. Under a second end-to-end, typically.

You don't train models, tune filters, or run a GPU. You ship hardware and an app; we handle the signal processing.

What a working integration looks like

1. You register a Device Model in the dashboard (e.g. xyz_1, a wristband with a PPG green sensor).
2. You mint an API key (raeh_…) for your account.
3. Your firmware connects to /stream/ingest and streams binary frames of samples at the sensor's native rate.
4. Your phone/web app connects to /stream/subscribe and receives JSON insights in real time.

That's it. No other endpoints are required for the critical path.

Who these docs are for

• Firmware / device engineers integrating /stream/ingest
• Mobile / web app engineers integrating /stream/subscribe
• Backend engineers provisioning keys and querying session history

Start here

• Quickstart: 5 minutes to a live insight
• Concepts: the vocabulary of account, device model, session, insight
• Guides: deeper walkthroughs of each piece
• Reference: WebSocket protocols, REST endpoints, binary frame format