LLMpad

Português do Brasil | English

A lightweight desktop app for chatting with local LLMs, built with Tauri + React.

Features

• 🪶 Lightweight - Uses Tauri (Rust) instead of Electron
• 💾 Persistence - Conversations saved in local SQLite
• 🔌 Compatible - Works with any OpenAI-compatible API (Ollama, LM Studio, etc.)
• 🖥️ Cross-platform - Mac, Linux and Windows
• 🤖 Model Manager - Create and manage custom Ollama models

Requirements

• Node.js (v18+)
• Rust
• Ollama (recommended for local execution)

Installation

# Install dependencies
npm install

# Run in development mode
npm run tauri dev

# Build for production
npm run tauri build

Configuration

1. Open the app
2. Click on "Settings" (⚙️)
3. Configure the API URL:
   • Ollama: http://localhost:11434/v1
   • LM Studio: http://localhost:1234/v1
   • OpenAI: https://api.openai.com/v1
4. Select the model from the dropdown
5. API Key is optional for local APIs

Managing Custom Models

Creating a new model

1. Create a .Modelfile in the models/ directory
2. Example (models/my-model.Modelfile):

   FROM llama3.2
   
   PARAMETER temperature 0.7
   PARAMETER num_ctx 2048
   
   SYSTEM """
   You are a helpful AI assistant.
   """
   

3. In the app, go to Settings > Manage Models
4. Click "Create in Ollama" next to your model

Recommended Base Models

• llama3.2 - Good balance of speed and quality
• phi3 - Fast and efficient model
• mistral - Great for non-English languages

Stack

• Frontend: React + TypeScript + TailwindCSS
• Backend: Rust + Tauri 2.0
• Database: SQLite (via rusqlite)
• Icons: Lucide React

License

MIT