This project is an interactive FAQ (Frequently Asked Questions) assistant built using Streamlit. It uses a dataset of questions and answers, which are preprocessed and converted into numerical vectors using a model called SentenceTransformer. When a user asks a question through the interface, the system quickly finds the most similar question from the dataset using a method called Annoy indexing. If a match is found, the corresponding answer is displayed; otherwise, the system uses the GPT-2 model to generate a relevant response. The entire application is designed to be user-friendly and can be embedded on websites to assist visitors with their queries.