AI-On-The-Edge-Cam: Esp32-S3 with PoE, SD, Camera

What is it?

This is an ESP32-S3 development board (16 MB flash, 8 MB RAM) that is compatible with the AI-on-the-Edge-Device framework. It integrates Ethernet with PoE, a micro-SD card slot, a camera with WS2812B back-lighting LEDs, a Stemma QT connector, comprehensive battery management, and a low-power sleep mode that draws just 20 µA.

Why did you make it?

It was created because, some time ago, I came across a great edge-AI framework called AI-On-The-Edge-Device and, after some discussion with the creators, I decided to build this custom hardware. The main motivation was that the ESP32-Cam, which is mostly used with the framework, lacks good power and battery management, camera back-lighting, and Ethernet/PoE support.

What makes it special?

• Ethernet with PoE
• Camera with 4 × WS2812B smart LEDs for back-lighting
• Micro-SD card slot
• 20 µA sleep current
• Integrated battery charging, protection, and measurement
• External antenna support
• Stemma QT I²C connector for external sensors
• Compatible with Ai-on-the-edge-device Esp32-s3 branch
• Schematics available via Datasheet tab

Each package contains:

• 1 × AI-on-the-edge-cam board

Product price includes 21% Czech VAT.

Other resourses

• Arduino IDE & Platformio examples
• Hackaday project
• Ai-On-The-Edge-Device framework
• Ai-On-The-Edge-Device Wiki
• Project discussion with Ai-On-The-Edge-Device developers

Warnings

• In V2.1.0, the battery-measurement circuit lacks a capacitor; I’ll fix this in future revisions. For now, if you need battery measurement, either solder a 0.1–1 µF capacitor between GPIO2 and GND (there’s a grounded M3 mounting hole nearby), or use software filtering. I’ve had good results with a mean-of-medians filter (compute medians from seven 41-sample windows, then average them).
• If you purchase pre-flashed version, don't treat it as a final device, It may have bugs because the software is in development. I tested all the basic features including WiFi&Lorawan&Ethernet. By purchasing it you support the developers of Ai-On-The-Edge-Device framework!