KEHIPI
KEHIPI Raspberry Pi AI HAT+ 26 TOPS Development Board
Add-on board with Hailo AI accelerator for Raspberry Pi 5 delivering 26 TOPS for high-performance AI tasks like robotics and home automation
- 26 TOPS accelerator runs larger neural networks faster
- Hailo-8 inference engine handles multiple AI models simultaneously
- PCIe Gen 3 link enables automatic detection on Raspberry Pi 5
- Native rpicam-apps support offloads post-processing to the NPU
- Accessible, power-efficient AI for robotics, security and automation
AI Accelerator for Raspberry Pi 5
The KEHIPI Raspberry Pi AI HAT+ is an add-on board that brings a built-in Hailo neural network inference accelerator to the Raspberry Pi 5. It provides an accessible, cost-effective, and power-efficient method to integrate high-performance AI capabilities. Target users include makers, developers, and engineers exploring applications such as process control, security, home automation, and robotics.
HAT Form Factor with PCIe Gen 3 Interface
The board adopts the standard HAT form factor, mounting directly onto the Raspberry Pi 5 GPIO header for a secure mechanical connection. It communicates with the host via the Raspberry Pi 5 PCIe Gen 3 interface, enabling high-throughput data transfer for inference workloads. When running an up-to-date Raspberry Pi OS image, the system automatically detects the on-board Hailo accelerator and makes the NPU available for AI computing tasks without manual driver configuration.
26 TOPS Neural Network Performance
This variant delivers 26 tera-operations per second (TOPS) of inference performance, built around the Hailo-8 accelerator. The higher throughput allows it to handle larger neural networks, run them more quickly, and more effectively execute multiple networks simultaneously compared to the 13 TOPS variant. The built-in rpicam-apps camera applications in Raspberry Pi OS natively support the AI module, automatically using the NPU to run compatible post-processing tasks.
أبرز المميزات
- 26 TOPS accelerator runs larger neural networks faster
- Hailo-8 inference engine handles multiple AI models simultaneously
- PCIe Gen 3 link enables automatic detection on Raspberry Pi 5
- Native rpicam-apps support offloads post-processing to the NPU
- Accessible, power-efficient AI for robotics, security and automation
المواصفات
| Brand | Generic |
|---|
أسئلة حول هذا المنتج
What PCIe generation does the HAT+ use and what bandwidth does that provide in practice?
The board communicates over the Raspberry Pi 5 PCIe Gen 3 interface, giving the Hailo-8 accelerator a dedicated high-speed lane for inference workloads.
What physical clearance does the HAT+ need on the Raspberry Pi 5?
It mounts directly onto the Pi 5 GPIO header and sits above the board, so allow enough vertical space for the HAT+ thickness and any attached camera module.
How is the HAT+ enabled in Raspberry Pi OS and what step is often missed?
Run an up-to-date Raspberry Pi OS image; the system auto-detects the Hailo accelerator, but users often forget to update the OS and firmware before first boot.
أيضًا في هذا القسم
عملاء قارنوا هذه المنتجات
نفس القسم، ونفس صفحة الدفع — ثماني عملات رقمية وقفل سعر لمدة 30 دقيقة.



