NeoEyes NE301 PoE Edge AI Camera

Product Overview

NeoEyes NE301 PoE is the wired-power version of CamThink’s STM32N6 edge AI camera. It is designed for installations where continuous operation and stable Ethernet are more important than battery runtime, such as outdoor monitoring points, smart facility inspection, equipment rooms, construction sites, and other unattended edge deployments.

The PoE version adds a PoE expansion board to the NE301 platform. It supplies device power and wired network connectivity through a standard Ethernet cable, while the camera still keeps the NE301 Web UI workflow for configuration, model management, stream settings, and data reporting. For teams evaluating the full platform, the NE301 product overview documents the STM32N6 MCU, modular camera design, expansion interfaces, IP67 enclosure, and open-source development ecosystem.

Inside the camera, the STM32N657L0H3 integrates an Arm Cortex-M55 core and Neural-ART accelerator for local visual inference. The factory model is a YOLOv8 Nano object detection model, and developers can replace model packages through the Web UI. NE301 can report inference results through MQTT/MQTTS or Webhook, push live streams through RTMP/RTSP, and connect to external triggers such as PIR sensors, radar, Alarm IO, or RS485 devices.

Key Features

PoE power and wired Ethernet

NE301 PoE uses IEEE 802.3af PoE to carry power and data through one Ethernet cable. This simplifies installation where PoE switches or injectors are already available, and it removes battery life as a deployment constraint for continuous capture, live streaming, and scheduled monitoring.

STM32N6 on-device AI

The camera runs inference locally on the STM32N6 Neural-ART accelerator, rated up to 600 GOPS / 0.6 TOPS. It supports TFLite Int8 model inference, with Web UI model upload for project-specific AI tasks. See the model training and deployment guide for the supported deployment workflow.

Web UI workflow for setup and debugging

The built-in Web UI supports live preview, model management, inference parameter tuning, media streaming settings, data reporting, storage management, firmware upgrade, and configuration import/export. The PoE version still uses the local NE301 WiFi AP for device setup, with access through 192.168.10.10.

Data reporting for system integration

NE301 can publish AI inference results and device data through MQTT/MQTTS, or send one-way HTTP(S) data through Webhook. For bidirectional control, MQTT supports reporting plus remote commands. See the MQTT data interaction guide for topic and JSON reporting details.

RTMP/RTSP video streaming

For video-based monitoring and AI result visualization, NE301 can push streams to an external server by RTMP or RTSP through the Web UI. The PoE version is especially suitable for this workload because wired Ethernet and continuous power reduce deployment risk. See the RTMP video streaming guide.

Outdoor and trigger-ready deployment

The PoE version is documented for outdoor waterproof, continuous capture, and streaming scenarios. It uses the NE301 IP67 enclosure and exposes Alarm and RS485 interfaces on the PoE board. PIR and other sensor-triggered workflows are supported, with setup guidance in the PIR sensor integration guide.

Lens Configuration