AI in every device!

supports easy and simple AI implementation.

AI adoption is becoming more active in various scenes. In recent years, it has also been expanding into the embedded device industry. Even in such scenes of "embedded AI", More easily, more simply. We will realize the implementation of AI that can be used in the existing development environment.
ファンレス 低消費電力
ailia SDK Free Trial VersionApplication for trial.
Support and Consultation.Contact
I'll answer any questions you have about Ainyaailia, meow.
Ai-nyan ChatBOT (automatic response)


Do you have any challenges like this?
I bought an AI-compatible embedded board.
I don't know where to start.
Embedded devices
Manuals and SDKs are complex and difficult.
Without changing the existing system.
I want to add AI.
If you use "ailia SDK", it is possible to smartly solve such issues.
With abundant models and samples ready, you can easily operate AI. In addition,
since "ailia" is pre-installed on devices with "ailia Ready",
there is no need for the hassle of setting up the environment.

It is a safe and comfortable AI development environment
where you can easily try out a diverse range of AI that can run on any platform,
evaluate it, and then proceed to actual development.


Introduction flow
"ailia SDK" × SoC is a one-stop solution.
There are various strengths in adopting it.
Installation of the high-speed AI inference framework "ailia SDK".
Not necessary for ailia Ready (= pre-installed) compatible devices.
I have prepared over 300 types of samples.
Choose the model you want to run from "ailia MODELS".
Even more convenient on Android. "ailia AI Showcase" as well providing.
ailia AI showcase is an application that allows users to easily try out various AI functions. With the deep learning framework "ailia SDK", it achieves fast inference by making the most of CPU, GPU, and NPU.
App Store
Google Play


Why is Ailia
Optimal for embedded devices?
NPU compatible
It is possible to run AI with low power consumption using int8. With this, it is now possible to embed AI even in embedded environments that had given up on adoption due to power consumption issues.
CPU and GPU compatible
We have optimized it for the mainstream Arm NEON in embedded devices, and the GPU in embedded devices can be used with cuDNN and Vulkan.
You can develop AI on a PC without specialized knowledge for embedded devices, because it provides the same API for both PCs and embedded devices.
Abundant model library
Just choose the AI you want to use from a wide range of model libraries without the need for relearning with the customer.


In embedded devices.
AI Utilization Scenarios.
Use case of surveillance camera
Embedded AI and surveillance camera market cloud have a high affinity, ailia runs AI on edge devices to reduce latency, reduce costs, and also suppress heat with Int8 support so that data transmission and server costs do not increase.
Use cases for automobiles.
Ailia is expected to support the prevention and improvement of dangerous driving such as distracted driving, speeding, and driver drowsiness, as well as daily peace of mind and safety in logistics and transportation infrastructure scenes, by supporting generic SoC and Android in automobiles and raising them to a higher level.
ailia Ready Verified System


Specialized staff. We will support you in return.
Support from introduction to development to use
Both email and meetings are possible.
Development bases and headquarters functions are both domestic.
Prompt and thorough follow-up.
I will do it (or) I will take care of it.

Frequently Asked Questions

Frequently Asked Questions
Regarding frequently asked questions about inquiries,
I will post the answer.
  • QWhat types of embedded SoCs do you support?
    AThe translation of the given text is: "Compatible models are Qualcomm QCS610, Mediatek Genio1200, NXP iMX8M Plus, and TI TDA4AL."
    "I plan to continue increasing in the future."
  • QWhat kind of embedded OS do you support?
    AIt depends on the SoC, but Linux and Android.
    Please contact us for more information.
  • QWhat format does the learning model support?
    AONNX is supported. ONNX supports only float models and can be used with CPU and GPU.
    TFLite supports float and int8 models, and can be used with CPU and NPU.
  • AWhat development languages are available?
    AC++ and Python.
    Furthermore, Unity can be used as a development tool.
  • QHow can I purchase a board with SoC installed?
    AYou will need to purchase from each SoC manufacturer or SoC board manufacturer's distributor.
    Please contact us as it is possible to introduce you from Accel.
  • QI would like to request development, is it possible?
    AIt is possible to propose from the Axel group or from partner companies.
    Please contact us for more details


Documents and sample programs, etc.
We have prepared it for you.
  • DocumentI will introduce the implementation method and the details of the API.
  • Support SlackIt is possible to ask questions via chat.
  • Support and Consultations.ContactPlease feel free to contact us.


Development and implementation support
Not only providing the library, but also...
Consulting, development, and implementation are also available.

AI development support service
"ailia WORKS"

Not only the provision of AI technology, but also through various AI development support, we propose the optimal problem solving to our customers.
We will meet all customer needs, such as implementing minutes of meetings and developing voice control functions.
  • Development of AI applications.
  • Development of server systems using cloud technology.
  • Development of AI model algorithm
  • Porting to embedded devices.
ailia SDK adopted enterprise list
ailia SDK Free Trial VersionApplication for trial
Also supports free individual consultations.Consultation on introduction

ailia SDK free trial version

ailia SDK
ailia AI Showcase
iOS versionApp Store
Android versionGoogle Play