QueryMate

QueryMate is a powerful, local command-line AI assistant built from scratch using C++ and the llama.cpp library. It brings the power of Large Language Models (LLMs) directly to your terminal, completely offline and portable.

Features

• 100% Local & Private: No data leaves your device.
• Context Aware: Maintains conversation history (short-term memory) for natural interactions.
• File Integration: Use @/path/to/file syntax to let the AI read and analyze local files.
• Multi-Model Support: Switch between different GGUF models (Llama 3, Gemma, Qwen, etc.) on the fly.
• Portable: Can be compiled and run from a USB drive on compatible machines.

Architecture

This project uses a modular Object-Oriented design:

• AssistantCore: The central controller managing the application flow.
• LlamaInteraction: Handles low-level inference using llama.cpp.
• MemoryManager: Manages conversation history and context window.
• CommandParser: Interprets user commands (e.g., /quit, /history, file references).
• UIUtils: Manages the CLI interface.

Building the Project

Requirements:

• CMake
• C++ Compiler (Clang/GCC)
• Make

mkdir build
cd build
cmake ..
make

Usage

1. Download a GGUF model and place it in the models/ directory.
2. Run the executable:

   ./build/chat

3. Chat naturally or use commands like /help.

Credits

Powered by llama.cpp.
Inspired by Gemini Cli.