GitHub Store
dineshsoudagar

local-llms-on-android

dineshsoudagar

An Android application that brings local LLM chat to your phone, fully offline, private, and fast. It supports ONNX-based Qwen models and LiteRT-based Qwen 3 and Gemma 4 models.

v1.6.0
Published at
19d ago
437
Stars
18,614
Download
Synced at
15d ago
android

Latest release

v1.6.0Sep 30, 2026, 02:51 PM
Downloads slow? Switch the source anytime above

README

Project Overview

Local LLMs on Android is an Android application that brings local LLM chat to your phone, fully offline, private, and fast. It supports ONNX-based Qwen models and LiteRT-based Qwen 3 and Gemma 4 models.

Key Features

  • Run quantized LLMs (Gemma, Qwen, LLaMA, etc.) fully offline
  • Dual backends: ONNX Runtime and LiteRT
  • Streaming response support
  • Local, private chat and question answering
  • Ready-to-install APK releases

Target Audience

  • Users who want to run LLMs offline on Android phones
  • Privacy-conscious users who don't want conversation data uploaded
  • Users with low-end to mid-range Android devices

Quick Start

  1. Download the APK for your model from GitHub Releases (e.g., Qwen3 0.6B LiteRT for low-end phones)
  2. Install on your Android device
  3. Open the app and start chatting offline

Core Advantages

  • Fully offline: Runs without network connection
  • Privacy protection: Data never leaves the device
  • Fast inference: Optimized with ONNX Runtime and LiteRT backends
  • Low barrier: Lightweight model versions available for low-end phones

System Requirements

  • OS: Android

Notes

  • The Qwen3 0.6B LiteRT version is best for low-end phones
  • Actual speed and memory usage depend on model size and device performance

Release history

Sort