Hermit

React Native/Expo mobile chat app with local LLM inference via llama.rn. Supports both remote OpenAI-compatible APIs and local on-device GGUF models.

Features

• Remote API and local on-device inference
• GGUF model import and management
• Per-model sampling parameters and system prompts
• Chat management with copy, edit, and regenerate
• Thinking/reasoning display (R1, XML, separator formats)
• Secure API key storage via expo-secure-store

Quick Start

npm install
npm start          # Expo dev server
npm run android    # Android
npm run ios        # iOS (macOS only)
npm run web        # Web

Setup

Remote API

1. Drawer > Remote API Configuration > Add config (name, base URL, API key)
2. Drawer > Model Configuration > Select API source > Add model config
3. Select model from header dropdown and chat

Local Model

1. Drawer > Local Model Configuration > Import .gguf file > Configure and test
2. Drawer > Model Configuration > Switch to Local Model source > Add model config
3. Select model from header dropdown and chat

Tech Stack

React Native 0.81 / Expo SDK 54 / expo-router 6 / expo-sqlite / llama.rn 0.12 / expo-secure-store

Docs

See docs/ for detailed documentation on architecture, components, database schema, and inference.

License

MIT