PrivateGPT Chatbot

Overview

PrivateGPT is a chatbot designed to answer questions based on your documents. It ensures privacy by processing everything locally.

Features

• Document ingestion and vector storage using Weaviate.
• Retrieval-based Q&A with LangChain and HuggingFace.
• Simple frontend using Streamlit.
• Backend powered by FastAPI.

Workflow

Setup Instructions

1. Clone the Repository

git clone https://github.com/your_username/privateGPT.git
cd privateGPT

2. Install Dependencies

conda create -n privategpt_env python=3.10 -y
conda activate privategpt_env
pip install -r requirements.txt

3. Prepare and process Documents Place your .pdf and .txt documents in the source_documents/ directory. Then run:

python ingest.py

This creates a local FAISS vectorstore for your documents.

4. Rub Chatbot Start the chatbot by running:

python app.py

5. Use the Web Interface Run the Streamlit UI:

streamlit run app_ui.py

Open the displayed URL in your browser to interact with the chatbot.