Secure Payment
Fast Delivery
Money Back Guarantee
24/7 Support
Quality Assurance
Free shipping on all orders . See details
Estimated between Oct 05, 2026 and Oct 07, 2026 Ships within 1-3 business day after payment clears. | SHOP Track Order | Ourlum Orders | Track Orders
30 days returns. 100% Money back guaranteed. See details
Best performance (INT8): 26TOPS
Brand Name: EASY EAI
Condition: New
High-concerned chemical: none
Operating Temperature: -40°C-80°C
Origin: Mainland China
Type: Module
Guangzhou EASY EAI Technology Co., Ltd.: An enterprise focusing on artificial intelligence IoT products and related solutions. The company's main products are AI main control boards, AI core boards, and AI edge computing boxes. At the same time, we provide customized services for hardware products and AI algorithms to accelerate customers’ product development and implementation. The customer service replies to hailo-8 to get the information.
Online documentation: https://www.easy-eai.com/document_details/19/615
Google Drive data:https://drive.google.com/drive/folders/1SSzqwXMYyCkkFb7CCpN1OpOVGKh2PMWe?usp=sharing
Business Email: sales.global@easy-eai.com
WhatsApp:+86 13450496436
model zoo: https://github.com/hailo-ai/hailo model zoo/tree/master
Code Routine:https://github.com/hailo-ai/Hailo-Application-Code-Examples
Shop OurLum with confidence—secure payments, fast shipping, easy returns, order tracking, and reliable support.
At OurLum.com, we don’t just promise quality—we guarantee it. Every product is carefully selected and tested to meet real-life needs, and your satisfaction is 100% backed by our no-hassle return policy. If it doesn’t work for you, we’ll make it right—fast. Shop confidently knowing you're getting trusted products, secure checkout, and real support when you need it. Because at OurLum, your peace of mind comes first
Simply FANTASTIC, arrived very well packaged in a custom Hailo box, worked perfectly with my N100 running Frigate for 6 cameras, with a temperature around 75ºC, running MobileNet v1.
Real deal! Identifies and performs as hailo-8 should. Frigate runs smoothly using yolov11s detector.
# Docker exec frigate hailortcli fw-control identify
Executing on device: 0000:04:00.0
Identifying board
Control protocol version: 2
Firmware version: 4.21.0 (release,app,extended context switch buffer)
Logger version: 0
Board name: hailo-8
Device architecture: hailo8
I checked, it works and is detected by the lspci command.
The module is genuine hailo-8 with 26 tops. Module when inserted into nvme m.2 slot (not just sata m.2) is recognized both in windows (hwinfo64) and linux (tried ubuntu). I have used it to build a designated nvr - inserted the module to a mini pc, installed ubuntu lts and drivers, frigate in docker and paired it to my home assistant. Works perfectlybwith frigate and yolov9s model. Much less false detections than with coral and now i can also run detection on higher resolution streams.
Shipping was fast and packaging good - module in sealed bag, well packed in bubble wrap and shipped in a plastic box. Price could be better, especially when compared to raspberry ai hat which is basically the same chip just not in m.2 forman. However, i wanted the m.2 version for the mini pc and compated to other such version this price is reasonable. Overall, i am satisfied with the product and the seller.
It arrived on time!
Everything is as described, managed to run it on TrueNas Scale 25.10.2.1 - Goldeye. There is no driver 21 for the newer version of TrueNas and it needs to be activated in Frigate.
Thankyou seller, the module has arrived in good order, and well packaged.
Great price, considering how difficult these are to find.
Works perfectly in my frigate setup.
Arrived in the uk in a week.
Installed it in RP5 via the M2 board. Everything works perfectly with Frigate, the objects are detected. I installed an additional heatsink on the board and am cooling everything with a fan, but there is still a slight amount of heat. RP5 16GB (8GB would be sufficient) + Hailo 8 can handle 10 FullHD + 2 4K cameras, CPU load on RP is 30-50% depending on traffic intensity, detector time is 30-40ms. I am very satisfied with the system. Intel+Nvidia would handle around 300W, but here it's no more than 15W. I'll be honest, I had a bit of a struggle with selecting the drivers and the Linux kernel, but only for those who aren't afraid of challenges and understand why it's necessary!
Great, I managed to compile the driver for DSM7.2 (xpenology), and it works perfectly with Fridge.
The order has been received. Thank you very much.
All good very good 👍
It works. But the textolite was broken off, so I had to buy an adapter.
