New open-source software lets developers run spiking neural network models on BrainChip’s AKD1500 across major microcontroller platforms.
On 18 August 2026, BrainChip and hardware partner Neuromorphyx released an open-source software stack and Arduino-compatible library designed to run on BrainChip’s AKD1500 ultra-low-power neuromorphic processor. The software is built around the compact BrainBoard1500 developer/evaluation board (in the Arduino Nicla form factor), manufactured by Neuromorphyx using BrainChip’s AKD1500 neural-network accelerator.
The bundle includes open-source Arduino Nicla drivers, board libraries, and an Akida Engine software integration path. The open-source adaptation path enables developers to interface the neuromorphic chip with microcontrollers and single-board computers, including Raspberry Pi, Espressif (ESP32), Seeed Studio/XIAO, STMicroelectronics, OpenMV, Adafruit, and SparkFun.
Developers can run Spiking Neural Network (SNN) models directly on real-time sensor streams via standard SPI/QSPI interfaces with I2C system controls and onboard model flash. The company supplies customer-ready open-source AI model repositories tailored for edge use cases such as keyword spotting, visual wake words, and human activity recognition.
“Partnering with Neuromorphyx puts the AKD1500 in the hands of developers in a form factor and software ecosystem they already know,” noted BrainChip CEO Sean Hehir on the partner-led release. The software release aims to democratise neuromorphic edge AI, allowing developers to execute, measure, and refine event-based spiking neural network (SNN) models within constrained power, size, and latency limits across robotics, industrial IoT, space, and defence applications.
