ChatterUI

Active
GitHub TypeScript AGPL-3.0

Description

ChatterUI is a native mobile chat frontend for LLMs that can run on-device models or connect to OpenAI, Claude, Ollama and many other APIs.

Key Features

  • On-device inference — runs gguf models directly on Android via cui-llama.rn over llama.cpp
  • Multi-API support — OpenAI, Claude, Cohere, OpenRouter, Mancer, Ollama, koboldcpp and more
  • Character card spec — compatible with the Character Card v2 format used by SillyTavern
  • Tunable sampling and prompts — Sampler fields and Instruct formatting are fully customizable
  • Offline and private — Local Mode keeps all conversations on device
  • Device TTS — uses the system text-to-speech engine to read replies aloud

Use Cases

💡 Running quantized gguf models on Android for offline AI roleplay
💡 Connecting to a self-hosted Ollama or koboldcpp backend for private conversations
💡 Migrating existing character cards from SillyTavern to mobile
💡 Daily chatting on a phone via OpenAI or Anthropic-compatible APIs
💡 Trying out LLM chat on devices without a GPU

Tags

chat-ui mobile react-native llama on-device

Categories

💬 Chat UI

Quick Start

# Prepare Android SDK (API 33+) and Java 17/21
# Clone the repo and install dependencies
git clone https://github.com/Vali-98/ChatterUI.git
cd ChatterUI
npm install

# Start an Android dev build
npx expo run:android

# After first launch:
# 1. Open Settings -> Local Mode and turn it on
# 2. Open Models -> Import Model and pick a gguf model
# 3. Load the model and start chatting from the Chat page
Visit GitHub View Docs

Related Projects