RunAnywhere Web SDK - LlamaCpp backend for on-device LLM/VLM inference

Expo module for on-device LLM inference via Apple Foundation Models (iOS) and Gemini Nano (Android)

On-device LLM inference for the browser and Node.js — text, vision, speech (TTS/STT), and embeddings on a native WebGPU engine. Offline, private, no API keys. React hooks, MCP, and Vercel AI SDK included.

React Native library for on-device LLM inference using llama.cpp. Part of Novastera CRM/ERP platform ecosystem.

LlamaCpp backend for RunAnywhere React Native SDK - GGUF model support for on-device LLM

On-device LLM inference for React Native with Metal (iOS) and Vulkan (Android) GPU acceleration

LiteRT-LM provider for @localmode - Google's on-device LLM inference engine with WebGPU/CPU support, runs .litertlm models (Gemma 4, Qwen3)

On-device LLM sessions for any browser — Chrome's Prompt API, WebLLM, or wllama, picked by feature detection, smallest model first.

On-device LLM inference for React Native — Gemma 4, Gemma 3, Phi-4, Mistral, Llama 3 and more via Google AI Edge LiteRT-LM, ML Kit AICore and MediaPipe

Unofficial TypeScript bindings for Apple's Foundation Models framework (Apple Intelligence on-device LLM)

On-device LLM inference for React Native. Run Qwen 2.5 models locally with Metal on iOS, CPU on Android. No cloud, no API key.

Web → native bridge SDK for the Dash Chat WebView host: on-device LLM / STT / TTS engines, a framework-agnostic chat engine, and deep-link navigation.

Expo module for Apple's on-device LLM (Foundation Models, iOS 26+). One-shot generation and multi-turn LanguageModelSession, no network or API key.

Natural language query parser for the Lemma protocol. Converts natural language into structured attribute queries using on-device LLM inference via [Transformers.js](https://huggingface.co/docs/transformers.js).

Expo module for Apple's Foundation Models (on-device LLM) and CoreML

On-device LLM + RAG + memory for Expo & React Native using MediaPipe

Dazzle SDK for React Native — embedded in-process database with HNSW vector search and a ChatAgent runtime for on-device LLM agents. Includes five swappable LLMClient adapters (LlamaCpp, LiteRT-LM, FoundationModels, OpenAI-compatible, Anthropic). Same lib

Velat — on-device LLM and RAG SDK. Placeholder. Full release Q3 2026. Visit https://velat.ai

GitHub OAuth Device authentication strategy for JavaScript

Detect device type and render your component according to it