The Oracle

Your personal know-it-all chatbot. The Oracle stores provided data and gives insights on everything involving that data.

Features

• Access to well-established Large Language Models (LLMs), such as Llama-3.3-70b and GPT-4o-mini;
• Simple and customizable interface;
• Support for YouTube and sites URLs, PDFs, CSVs and TXTs;
• Data memory.

Installation

Since this project uses Streamlit as framework, make sure to run streamlit run .\web.py to have access to the tool.

Option 1: Download ZIP

In the repository page, look for the green Code button and press it. Then, go to the Download ZIP option and click it.

A copy of the repository should start downloading with .zip format. Extract the folder on any directory, open it with the terminal and run:

$ pip install -r .\requirements.txt

Note

You can run it on your own machine, but it's preferable to run on a virtual environment

Finally, you can run the Streamlit initiation command and start using The Oracle.

Option 2: From source

$ git clone git@github.com:gabrieldpbarros/Oracle-Project.git
$ cd Oracle-Project/
$ pip install -r .\requirements.txt
$ streamlit run .\web.py

Usage

Accessing the web app, the user needs to select a model in the Model Selection tab and insert the corresponding API Key

Important

The user must request an API Key from the provider externally. If you want to run a Groq model, you must have access to a Groq API key, for example.

Then, you can return to the Archive Upload tab and select whichever type of upload as you desire.

With all this completed, now you can run the model and start chatting with it.