Nexa SDK

What is Nexa SDK?

Nexa SDK is a powerful toolkit designed to enable developers to deploy various AI models, including LLMs, multimodal, ASR, and TTS, across a wide range of devices such as PCs, mobile phones, automotive systems, and IoT devices. It is optimized for performance and privacy, ensuring that applications can run efficiently on NPU, GPU, and CPU architectures.

This SDK provides a seamless experience for integrating advanced AI capabilities into applications, making it suitable for both developers and businesses looking to leverage on-device AI technology. With Nexa SDK, users can expect fast, production-ready solutions that enhance the functionality of their devices.

Supports deployment on multiple platforms: mobile, PC, automotive, and IoT.

Optimized for various hardware architectures: NPU, GPU, and CPU.

Enables fast and private AI model inference.

Nexa SDK Features

Nexa SDK is designed to facilitate the deployment of various AI models, including LLMs, multimodal, ASR, and TTS models, across a wide range of devices such as mobile, PC, automotive, and IoT. It ensures fast, private, and production-ready performance on different processing units, including NPU, GPU, and CPU.

Key features and capabilities of Nexa SDK include:

Support for multiple AI model types, enabling versatile applications.

Optimized for various hardware platforms, ensuring efficient performance.

Designed for on-device AI, enhancing privacy and reducing latency.

Production-ready solutions that can be easily integrated into existing systems.

Why Nexa SDK?

The Nexa SDK offers a powerful solution for developers looking to deploy advanced AI models, including LLMs, multimodal, ASR, and TTS, across a variety of platforms such as mobile, PC, automotive, and IoT devices. Its architecture is designed to be fast, private, and production-ready, ensuring that applications can leverage the full potential of NPU, GPU, and CPU resources.

By utilizing the Nexa SDK, developers can benefit from:

Seamless deployment of AI models across multiple device types.

Enhanced performance with optimized inference on various hardware architectures.

Robust support for both on-device AI and edge computing applications.

Improved privacy and security by processing data locally.

How to Use Nexa SDK

To get started with the Nexa SDK, first, you need to create a free account at[@portabletext/react] Unknown block type "span", specify a component for it in the `components.types` prop. Once you have your account, activate your device using your access token with the command:nexa config set license '<access_token>'. This step is crucial for enabling the SDK to function properly on your device.

After activation, you can run your desired model effortlessly. For instance, to execute a model on a Qualcomm NPU, simply use the command:nexa infer NexaAI/phi4-mini-npu-turbo. This streamlined process allows you to leverage the powerful capabilities of the Nexa SDK with minimal setup.

Lightweight yet capable: strong reasoning (math/logic) in a compact footprint.

Fast, private, and production-ready on NPU, GPU, and CPU.

Supports a variety of models including LLMs, multimodal, ASR, and TTS.

Ready to see what Nexa SDK can do for you?[@portabletext/react] Unknown block type "span", specify a component for it in the `components.types` propand experience the benefits firsthand.