No products in the cart.
NeoEyes NE301 Edge AI Camera
Price range: $199.90 through $258.00
1. On-Device Vision AI for Fast PoC and Validation
Run object detection, motion analysis, and pose estimation fully on-device, and no cloud required. Built on STM32N6 with integrated NPU, delivering up to 0.6 TOPS for real-time vision validation and demos.
2.Designed to Prototype, Debug, and Iterate
Deploy models, tune parameters, and test logic via Web UI or MQTT without rebuilding firmware. Ideal for fast PoC cycles, open-source projects, teaching labs, and internal technical evaluations.
3.Open, Controllable, and Developer-Friendly
Open-source firmware and SDK with support for STM32 Cube.AI, YOLO, MobileNet, and EfficientNet. Full access to inference logs and local storage via SD card.
4.Expandable Interfaces for Real Projects
Supports MIPI CSI and USB cameras, RS-485, GPIO, I²C, UART, and audio interfaces. Modular design enables expansion with PoE, LTE Cat.1, and external sensors for custom system integration.
5.Ready Beyond Demos and Bench Tests
IP67-rated enclosure with wide operating temperature -20 °C to 50 °C. Multiple power options including battery, USB-C, and PoE, making it suitable for pilots, outdoor testing, and small-scale deployments.
Wi-Fi included in all NE301 variants.
$20.00 OFF
EARLYBIRD-NE301
NE301 Early Bird $20
Applicable on selected products
Validate Vision AI Ideas — Fast
Built For Event-Driven Vision
Sleeps most of the time. Wakes only when events happen
Sleep
Standby
Performance
Event wake-up via MQTT / RTC / Bluetooth
Built For Event-Driven Vision
Sleeps most of the time. Wakes only when events happen
Designed to
Run Unattended
— For Months or Years
-
IP67
-
−20 °C to 50 °C
-
Multiple mounting options
-
Power Options: 4× AA battery、USB-C、PoE
NE301 WiFi Version
2.0-Year
Estimated Lifespan
NE301 LTE Cat.1 Version
1.5-Year
Estimated Lifespan
Test Conditions
4x AA Batteries
10 Shots / Day
Basic Mode
7 μA Deep Sleep
How long can NE301 run on batteries?
Battery life varies by configuration and usage.
For low-frequency event capture, NE301 is optimized for months to years deployments, not
days or
weeks.
NE301 System Variants
-
NE301 Wi-Fi Base Version
Best for Indoor or networked environments with existing Wi-FiConnectivity Wi-Fi (2.4 GHz) + BluetoothPower Low-power, battery-friendlyMobility Medium(Wi-Fi range dependent)Choose this if Already have Wi-Fi and want the simplest, lowest-cost setup*Wi-Fi included in all NE301 variants.
-
NE301 Cat.1 Version
Best for Remote, outdoor, or infrastructure-free deploymentsConnectivity 4G LTE Cat-1 (Data only)Power Low-power, optimized for long standbyMobility High (cellular coverage dependent)Choose this if Operate without local network access -
NE301 PoE Version
Best for Fixed installations with continuous operationConnectivity EthernetPower Always-on, full-performanceMobility Low (wired installation)Choose this if need stable connectivity and continuous AI or video workloadsComing Late March
Expandable When Needed
Start simple. Expand only when your project grows.
Motherboard (front)
Motherboard (back)
Overview
Technical Specifications
AI Processor & Memory
| MCU Core | Arm Cortex-M55 @ 800 MHz with Arm Helium (M-Profile Vector Extension) |
| NPU | Neural-ART™ Accelerator @ 1 GHz |
| AI Performance | Up to 0.6 TOPS |
| SRAM | 4.2 MB |
| HyperFlash | 128 MB |
| PSRAM | 64 MB |
Power & Efficiency
| NPU Efficiency | Up to 3 TOPS/W |
| Standby Current | ~7 μA (low-power mode) |
| Boot Time | Microsecond-level boot |
| Wake-up Time | Millisecond-level wake-up |
| Cooling | Passive (fanless) |
| Power Input | USB Type-C / Battery / PoE (variant dependent) |
| Power Ctrl Unit | The STM32U073Kx further reduces N6 power consumption through independent, refined power management. |
AI Software Support
| Pre-installed Analytics | Human detection & recognition |
| Model Compatibility | Compatible with lightweight CNN-based vision architectures, including YOLO-family, SSD-based detectors, MobileNet-class and ResNet-class models |
| Audio & Time-Series | YamNet, MiniResNet, Human Activity Recognition (HAR) |
| Custom Models | Supports INT8 TFLite models (BYOM) |
Imaging System
| Image Sensor | 1/2.9" Progressive Scan CMOS |
| Camera Module | 4 MP OS04C10 |
| Max Resolution | 2688 × 1520 (4 MP) |
| Video Encoding | H.264 hardware encoder, JPEG codec |
| Frame Rate | Up to 1080p @ 30 fps |
Lens Options
| 2.5 mm | 3.0 mm | 6.0 mm | |
|---|---|---|---|
| HFOV | 137° | 88° | 51° |
| Aperture | F2.4 | F2.5 | F2.2 |
| Distortion | < -33% | < 1% | < 1.6% |
Connectivity & Protocol
| Wireless | Wi-Fi 6 (802.11ax), Bluetooth 5.4 |
| Cellular | LTE Cat.1 (variant dependent) |
| PoE | Supported (variant dependent) |
| Protocol Stack | Based on LwIP and MbedTLS |
| Protocol Support | HTTP / HTTPS MQTT / MQTTS |
Interfaces & Expansion (Mainboard)
| Alarm Interface | 1 × Alarm In, 1 × Alarm Out (supports 4-pin smart PIR) |
| Storage | Built-in MicroSD / SDHC / SDXC slot, up to 2 TB |
| Multi-function Button | Reset / Wake / Capture |
| Lens Interface | 1× 4-pin USB, 1× MIPI CSI-2 |
| I/O Interfaces | 1× UART, 1× RS485, 1× I²C, 2× GPIO, 2× GND, 1× 3.3 V / 5 V (controllable) |
| Debug / Power | 1× USB Type-C, 1× 4-pin UART |
| Audio I/O | 1× Audio-In, 1× Audio-Out |
| Communication Expansion | 12-pin + 16-pin expansion IOs for communication modules |
| Power Interface | 2-pin connector (Battery / USB Type-C) |
| Sensor Expansion | Supports PIR, radar, temperature/humidity, and other sensors via motherboard I/O |
Environmental & Certification
| Operating Temperature | -20 °C to 50 °C |
| Operating Humidity | < 95% RH (non-condensing) |
| Protection Rating | IP67 |
| Dimensions | 77 × 77 × 48 mm |
| Certifications | CE, FCC |
-
How to assemble CamThink NeoEyes NE301 camera in a simple way
-
How to Update AI Camera Firmware CamThink NeoEyes NE301
-
AI camera (CamThink NeoEyes NE301)
FAQ
- Yes. The NE301 firmware is fully open-source and supports secondary development. You can access the source code and documentation on CamThink's GitHub.
- For hardware customization tailored to your specific needs, please contact us at sales@camthink.ai
NE301 currently supports YOLOv8. The device utilizes the N6 architecture as its model adaptation foundation. We are continuously expanding the library of compatible models based on the N6 core. For a broader list of compatible models, please refer to the STM32 Model Zoo - STMicroelectronics.
We offer professional model customization services for specialized industry scenarios. Feel free to reach out for technical consultation at support@camthink.ai
No.
The Cat-1 version requires a user-provided SIM card and cellular plan.
- Wi-Fi version relies on existing local Wi-Fi infrastructure and offers the lowest deployment cost in networked environments.
- Cat-1 version uses cellular connectivity and is designed for remote or infrastructure-free deployments.
Both versions share the same event-triggered AI vision core and low-power architecture.
You may also like…
-

NeoEyes NE101 Modular Sensing Camera
Price range: $69.00 through $112.00Select options This product has multiple variants. The options may be chosen on the product page -

NeoEyes NE101 Dev Kit – Modular Vision Camera
$59.90



















