Mayreau
Ultra-low power AI-ready platform with 8 sensors, STM32L4R9 MCU at 120 MHz, and Telit WE310F5-I WiFi/BLE 5.0 module. Raspberry Pi Model A form factor with USB-C. Open-source hardware.
Overview
Ultra-low power board designed for AI applications. Built around the STM32L4R9AII6 ARM Cortex-M4 with FPU running at 120 MHz and delivering 150 DMIPS. Equipped with 2048 KB Flash and 640 KB SRAM for demanding edge inference workloads.
The Telit WE310F5-I module provides WiFi 802.11 b/g/n and BLE 5.0 connectivity. Raspberry Pi Model A form factor with full expansion connector compatibility. USB-C for data and power. Compatible with 3x AA batteries. Open-source hardware. Follow development on Hackaday.io.
Integrated Sensors (8)
- BME688 — Environmental: temperature, humidity, pressure, VOC AI
- BHI160B — 6-DOF IMU: accelerometer + gyroscope
- BMM150 — 3-axis magnetometer
- IMP34DT05 — Digital MEMS audio
- IMP23ABSU — Analog audio + ultrasound
- TSL25911FN — Ambient light
- VL53L3CX — 8x8 multitarget Time-of-Flight proximity
- AMG8833 — 8x8 infrared thermal sensor array
Key Features
- STM32L4R9 150 DMIPS MCU
- Telit WE310F5-I WiFi + BLE 5.0
- 8 integrated sensors
- Raspberry Pi Model A form factor
- USB-C connector
- 3x AA battery compatible
- Open-source hardware
- Full RPi expansion connector
Development
Develop firmware using STM32CubeIDE. Suitable for edge AI inference with STM32Cube.AI, enabling on-device machine learning models optimized for the Cortex-M4 architecture.
8-sensor fusion applications: environmental monitoring, indoor positioning, occupancy detection, predictive maintenance.