RunanywhereAI/runanywhere-sdks

Provides unified SDKs across Swift, Kotlin, Web, React Native, and Flutter for running quantized LLMs, speech-to-text (Whisper), text-to-speech, and vision models entirely on-device using ONNX and llama.cpp runtimes. The architecture abstracts platform-specific inference engines behind a consistent API, enabling developers to build voice assistants and multimodal AI features without cloud dependencies. Includes pre-built sample apps demonstrating STT→LLM→TTS pipelines and image generation on mobile and web platforms.

10,245 stars. Actively maintained with 33 commits in the last 30 days.