ailia SDK

Getting Started

Choose your platform and follow three steps to run your first inference.

Install

Install the ailia Python package from PyPI. Python 3.6 or later is required.

pip3 install ailia

View on PyPI

Run a Sample

ailia-models ships 400+ ready-to-run inference scripts. Clone once, install the shared requirements, then cd into any model folder and run — pass -v 0 for webcam input, -i image.png for a still, or skip the flag to use the bundled demo input. python3 launcher.py at the repo root opens a GUI that browses every model.

git clone https://github.com/ailia-ai/ailia-models.git
cd ailia-models
pip3 install -r requirements.txt
cd object_detection/yolox
python3 yolox.py -v 0

Sample Repository (400+ models)

Clone Samples

Clone the C++ sample repository and initialize submodules. The ailia-sdk-cpp binding is included as a submodule.

git clone https://github.com/ailia-ai/ailia-models-cpp.git
cd ailia-models-cpp
git submodule init
git submodule update

Sample Repository Binding

Build & Run

Download a one-month evaluation license, install dependencies (CMake and OpenCV), build with CMake, and run a sample. Xcode and Android NDK starter projects are also available.

# Fetch a one-month evaluation license
cd ailia
python3 download_license.py
cd ..

# macOS
brew install cmake opencv
# Linux: apt install cmake libopencv-dev
# Windows: install CMake and Visual Studio,
#   then set OpenCV_DIR to your OpenCV build path

cmake .
cmake --build .
cd object_detection/yolox
./yolox.sh    # use yolox.bat on Windows

Xcode Project Android NDK Project

Install via UPM

For your own project, open Window > Package Manager in Unity, click + > Add package from git URL, and enter the binding URL below.

https://github.com/ailia-ai/ailia-sdk-unity.git

Binding

Run a Sample

Clone the Unity sample repository and open it in the Unity Editor (2021.3.10f1 or later). The SDK is downloaded automatically through Package Manager. Open a scene such as AXIP/AILIA-MODELS/FaceDetection/FaceDetectionSample.unity and press Play.

git clone https://github.com/ailia-ai/ailia-models-unity.git

Sample Repository

Add to pubspec

For your own project, add ailia as a git dependency in pubspec.yaml, then run flutter pub get. Flutter 3.19.6 or later is required. On macOS, set com.apple.security.app-sandbox to false in macos/Runner/Release.entitlements and Debug.entitlements.

dependencies:
  ailia:
    git:
      url: https://github.com/ailia-ai/ailia-sdk-flutter.git
      ref: main

Binding

Run a Sample

Clone the Flutter sample repository, open it in VSCode, and run flutter pub get. The SDK is downloaded automatically via pubspec.yaml. Pick a model and tap the run button.

git clone https://github.com/ailia-ai/ailia-models-flutter.git
cd ailia-models-flutter
flutter pub get
flutter run

Sample Repository

Clone Binding

For your own project, clone the JNI binding repository and add it to your Android Studio project.

git clone https://github.com/ailia-ai/ailia-sdk-jni.git

Binding

Run a Sample

Clone a sample project, initialize submodules, and open it in Android Studio. Both Kotlin and Java starter projects are available. Android Studio 2025.1.3 or later is required for the Kotlin samples.

git clone https://github.com/ailia-ai/ailia-models-kotlin.git
cd ailia-models-kotlin
git submodule update --init --recursive

Kotlin Samples Java Samples

Use the API in Your Project

Minimal examples for loading an ONNX model and running inference.

import ailia
import numpy as np

# stream=None lets ailia derive the graph from the .onnx file
net = ailia.Net(stream=None, weight="model.onnx")

input_tensor = np.zeros((1, 3, 224, 224), dtype=np.float32)
output = net.run(input_tensor)
print(output[0].shape)

#include "ailia.h"
#include <vector>

AILIANetwork* net;
ailiaCreate(&net, AILIA_ENVIRONMENT_ID_AUTO, AILIA_MULTITHREAD_AUTO);
ailiaOpenWeightFileA(net, "model.onnx");   // ONNX is self-contained

std::vector<float> input(1 * 3 * 224 * 224, 0.0f);
std::vector<float> output(1000);           // sized to the model's output
ailiaPredict(net,
             output.data(), (unsigned)(output.size() * sizeof(float)),
             input.data(),  (unsigned)(input.size()  * sizeof(float)));

ailiaDestroy(net);

using ailia;

var model = new AiliaModel();
model.OpenFile(null, "model.onnx");  // null prototxt for ONNX

var input = new float[1 * 3 * 224 * 224];
var output = new float[1000];        // sized to the model's output
model.Predict(output, input);

import 'package:ailia/ailia.dart';
import 'dart:typed_data';

final model = AiliaModel();
await model.openFile(streamPath: null, weightPath: 'model.onnx');

final input = Float32List(1 * 3 * 224 * 224);
final output = await model.predict(input);
print(output.length);

import axip.ailia.AiliaModel

val model = AiliaModel()
model.openFile(null, "model.onnx")   // null prototxt for ONNX

val input = FloatArray(1 * 3 * 224 * 224)
val output = model.predict(input)
println(output.size)

FAQ

Common questions from first-time ailia SDK users.

What is the difference between the evaluation license and a production license?

The evaluation license is downloaded automatically at runtime for Python, Unity, Flutter, and JNI, and is valid for one month for C++. It is intended for development and trial use only.

For commercial deployment, redistribution, or longer-term use, request a production license. See the ailia license terms for details.

How do I switch between CPU and GPU?

Pass an env_id to ailia.Net(). List available environments (CPU, CUDA, Metal, Vulkan, MKL, NNAPI) with ailia.get_environment_list(), then select the one you want.

By default, ailia chooses the fastest available environment for your platform.

Where are model files stored after I download them?

Sample scripts in ailia-models download .onnx and .onnx.prototxt files into the model's own directory the first time you run them. Subsequent runs reuse the cached files.

For ailia Speech and ailia Voice, models are downloaded into ./models/ by default, configurable via initialize_model(model_path=...).

Can I run ailia SDK offline?

Yes, after the first run. The evaluation license and any auto-downloaded model files require an internet connection on first use; once cached, subsequent inference works offline.

For C++, the license is fetched once via download_license.py from the binding repository.

How do I convert a PyTorch or TensorFlow model to ONNX for ailia?

Export your model to ONNX with the framework's standard tooling (torch.onnx.export, tf2onnx, etc.), then generate the matching .onnx.prototxt using the script bundled with ailia SDK.

The model conversion tutorial walks through the process.

Where can I get help?

For bug reports and questions on the sample repositories, open an issue on the relevant GitHub repo. For SDK licensing and commercial inquiries, contact ailia Inc. directly.

Getting Started

Install

Run a Sample

Clone Samples

Build & Run

Install via UPM

Run a Sample

Add to pubspec

Run a Sample

Clone Binding

Run a Sample

System Requirements

Operating Systems

Languages

GPU Acceleration

Model Formats

Use the API in Your Project

API Reference by Platform

Python

C++

Unity

Flutter

JNI

FAQ

Materials