boot-ai 🤖

A custom-built AI Agent capable of interacting with the local file system and executing Python code. This project was developed as part of the Build an AI Agent in Python module on Boot.dev.

The agent uses Google Gemini as its "brain" and is equipped with a suite of custom Python tools to bridge the gap between LLM reasoning and local system execution.

---

🛠️ Features

• Autonomous Task Execution: Uses Gemini's function-calling capabilities to choose the right tool for a given prompt.
• File System Operations: Read, write, and manage local files programmatically (see functions/).
• Dynamic Code Execution: A built-in tool to run Python scripts on the fly and capture output.
• Modern Tooling: Managed with uv for lightning-fast dependency management and reproducible environments.

---

🏗️ Project Structure

The project is organized into modular tools and a robust test suite:

• main.py: The entry point for the AI agent loop.
• functions/: Logic for the agent's core tools (get_files, get_file_content, write_files).
• calculator/: A command-line calculator app that the AI agent can work on.

---

🚀 Getting Started

Prerequisites

• Python 3.11+
• uv installed
• A Google Gemini API Key

Installation & Setup

1. Clone the repository:

   git clone [https://github.com/RussellGLarkin/boot-ai.git](https://github.com/RussellGLarkin/boot-ai.git)
   cd boot-ai
   

2. Install dependencies:

   uv sync
   

3. Configure environment: Ensure your Gemini API key is accessible via a .env file