RunanywhereAI/runanywhere-sdks

Production ready toolkit to run AI locally

What it solves

RunAnywhere provides a unified way for developers to integrate on-device AI capabilities—including LLMs, vision models, speech-to-text, and text-to-speech—into applications across multiple platforms without relying on cloud infrastructure. This eliminates latency, improves privacy by keeping data on the device, and enables offline functionality.

How it works

The project uses a shared C/C++ core (runanywhere-commons) with a C ABI, providing thin SDKs for iOS, Android, Flutter, React Native, and Web. It employs a plugin registry to automatically route inference to the most efficient engine available on the specific hardware, such as Core ML for Apple devices, WebGPU for browsers, llama.cpp as a general fallback, and a specialized runtime called QHexRT for Qualcomm Snapdragon Hexagon NPUs.

Who it’s for

Mobile and web developers who want to implement local AI features like chat, image understanding, and voice assistants without managing separate low-level inference engines for every platform.

Highlights

  • Multi-platform support: Single API for Swift, Kotlin, Flutter, React Native, and Web.
  • Hardware acceleration: Specialized NPU acceleration via QHexRT for Snapdragon devices, supporting MoE and hybrid-attention models.
  • Comprehensive AI suite: Supports text generation, vision-language models (VLM), speech-to-text (ASR), and text-to-speech (TTS).
  • Integrated pipeline: Ability to create full voice assistants by chaining STT, LLM, and TTS locally.
  • Prebuilt model bundles: Optimized model bundles available on Hugging Face for easy deployment.