testing commits

Fireball - Your Personal AI Assistant

Welcome to Fireball, your very own Jarvis-inspired AI assistant! This project aims to create a powerful, extensible AI system that can understand and respond to your questions while gradually building up capabilities through tools and worker agents.

Project Overview

Fireball is designed to be a personal AI assistant that:

• Connects to local LLMs for natural language understanding and generation
• Responds to user questions in a conversational manner
• Scales from simple question answering to complex multi-agent workflows
• Provides a foundation for building sophisticated AI tools and systems

---

🧠 Local LLM Assistant with UV

A simple yet powerful local AI assistant that leverages a locally hosted LLM and integrates tools such as weather lookup, task management, and folder search. Built using Python, with package management handled via UV.

📌 Overview

This project allows you to interact with a local large language model (LLM) through a terminal-based interface. It supports tool integration for tasks like fetching current weather, managing tasks, or locating folders on your machine.

Features

• ✅ Communicates with local LLM via OpenAI-compatible API
• ✅ Tool support: get_weather, find_folder, add_task
• ✅ Easy-to-use CLI interface with rich formatting
• ✅ Uses UV for fast, modern package management
• ✅ Supports multiple local LLM providers (Ollama, LM Studio, etc.)

🛠️ Prerequisites

Before running this project, make sure you have:

• A local LLM server running with OpenAI-compatible API:
  • Ollama - Recommended for ease of use
  • LM Studio - GUI-based model management
  • Or any other OpenAI-compatible local LLM server
• UV installed for package management

📦 Installation

1. Clone or download this repository.
2. Install dependencies using UV:

uv sync

This command will install all project dependencies listed in pyproject.toml.

3. Ensure your local LLM server is running:

   • Ollama: ollama serve (default: http://localhost:11434)
   • LM Studio: Start the server in LM Studio (default: http://127.0.0.1:1234)

4. Update the API endpoint in your configuration if needed to match your LLM server.

▶️ Usage

Run the application using:

uv run main.py

Then, you can interact with the LLM by typing prompts directly into the terminal.

Example Tools

You can use the following tools in your prompts:

• get_weather: Get current weather for a location.
  • Example prompt: "What's the weather like in London?"
• find_folder: Find folders matching a name.
  • Example prompt: "Find the devin folder"
• add_task: Add a task to the database with various details.
  • Example prompt: "Add a task called task1 with priority high"

🧪 Development

To add new tools or modify existing ones:

1. Add your function to tools.py.
2. Register it in ToolExecutor class under self.tools.
3. Update the system prompt in config.py if needed.

🔧 Configuration

The application supports different local LLM providers through configuration:

• Default configuration works with Ollama out of the box
• Easily adaptable to LM Studio, vLLM, or other OpenAI-compatible servers
• Modify API endpoints and model names in the configuration as needed

📄 License

This project is licensed under the GNU General Public License v3.0 – see the LICENSE file for details.

---

Roadmap

Phase 1: Core Functionality

• Basic LLM connection and response
• Conversation history management
• Enhanced response formatting

Phase 2: Tool Integration

• API capabilities
• File system operations
• Database interactions

License

GNU General Public License v3.0 - Non-commercial / Not for profit

---