🌍 Travel Guide

This is a practical project of an AI Travel Itinerary Generator that uses 100% local Artificial Intelligence. Powered by the LangChain framework, the system takes a destination and travel style to create a detailed, personalized itinerary.

🚀 Features

• Local Processing: Runs the Llama-3.2-1B model via GPT4All (100% privacy, no API keys required).
• Prompt Engineering: Optimized English templates for better logical reasoning, with output tailored to the user's needs.
• LCEL Architecture: Built using LangChain Expression Language for a clean and scalable data pipeline.

🛠️ Tech Stack

• Python 3.10+
• LangChain (Orchestration Framework)
• GPT4All (Backend for local model execution)
• Llama-3.2-1B-Instruct (Meta's lightweight LLM)

📋 Prerequisites

Before running the project, you must download the model using the GPT4All software and note the absolute path to the .gguf file.

Install the required libraries:

pip install langchain langchain-community gpt4all

⚙️ Configuration

In the main script, ensure you update the model path to the correct directory on your machine:

model_path = r"C:\Users\YOUR_USER\...\Llama-3.2-1B-Instruct-Q4_0.gguf"

🖥️ How to Use

The main script uses an input dictionary to define the trip details:

input_data = {
    "destination": "Rio de Janeiro",
    "days": "3",
    "style": "budget-friendly and cultural"
}

To run the generator:

python main.py

🧠 Project Structure

1. Model (LLM): Loading Llama-3.2 via the GPT4All integration.
2. Prompt Template: A structured instruction set that defines the AI's behavior as an expert travel guide.
3. Chain: Connecting the prompt and the model using the pipe (|) operator.