🚗 Vehicle Information System

AI-Powered CLI for Vehicle Data Management

---

📌 Overview

The Vehicle Information System is a command-line application that uses AI agents to interact with users, collect vehicle data (brand, model, fuel type, etc.), and retrieve matching vehicles from a database. It leverages Model Context Protocol (MCP) for structured client-server communication and LangChain for natural language processing.

✨ Key Features

✅ AI-Powered Queries – Ask questions naturally (e.g., "Show me red Toyotas from 2020")
✅ Multi-Language Support – Works in English & Portuguese
✅ Rich Terminal UI – Formatted output with Typer + Rich
✅ Database Abstraction – SQLAlchemy models with Faker-based seeding
✅ MCP Protocol – Standardized client-server communication

---

⚙️ Requirements

Before running, ensure you have:

• Python 3.12
• Ollama (for local LLM inference)
• Pipenv (for dependency management)

📦 Installation

1. Clone the repository

   git clone https://github.com/eacassecasse/c2s-virtual-agents.git
   cd c2s-virtual-agents
   

2. Set up a virtual environment

   pipenv shell
   pipenv install
   

3. Install Ollama & pull a model

   curl -fsSL https://ollama.com/install.sh | sh
   ollama pull llama3.2  # or mistral, gemma, etc.
   

4. Seed the database (optional)

   python -m scripts.seed
   

---

🚀 Usage

1. Start MCP Server

python -m app.server.server  # Runs on port 8000

2. Interactive Mode (AI Agent)

Start a natural-language conversation:

python -m app.cli.console

Example Queries:

• Show me blue Hondas
• List electric vehicles from 2022
• Mostrar carros com placa ABC-1234 (Portuguese)

---

---

🛠️ Development

Project Structure

c2s-virtual-agents/
├── app/
│   ├── cli/              # CLI commands & prompts
│   ├── core/              # Core features like the AI Agent
|   |-- db/                # Contains database related files
│   ├── models/           # SQLAlchemy models
│   └── server/            # MCP protocol logic
|   |-- storage/           # Storage engine management
├── scripts/
│   └── seed.py           # Fake data generation
├── tests/                # Unit tests
└── README.md
|__ .env                   # Environment variables

Testing

Run unit tests:

python -m unittest discover

Contributing

1. Fork the repo
2. Create a feature branch (git checkout -b feature/new-feature)
3. Commit changes (git commit -m "feat: add new filter option")
4. Push to branch (git push origin feature/new-feature)
5. Open a Pull Request

---

📜 License

MIT © Edmilson Cassecasse

---

Why Use This?

🔹 For Developers: Clean architecture with MCP for scalable client-server interactions
🔹 For Data Teams: Easily query & filter vehicle datasets
🔹 For AI Enthusiasts: LangChain + Ollama integration for NLP-powered search

Try it now and explore vehicles like never before! 🚀