-
-
Asking question from pdf
-
Menu for uploading pdf files
-
Can upload multiple pdfs
Inspiration
My inspiration for Pdf Answerer is from the difficultness of manually searching through lengthy PDF files or switching between multiple pdfs. I wanted to create a way to make it easier to ask questions and get accurate answers using Gemini AI API without any confusion.
What it does
It extracts text from PDF documents, process it, and answers user's question based on the content of those PDFs. User can upload one or more pdfs and can ask his doubts.
How we built it
I have built it using Python libraries like PyPDF2 for PDF text extraction, FAISS for efficient text indexing and similarity search and Google's generative AI model Gemini, for question answering (using prompt template).
Challenges we ran into
One of the main challenges that I faced was optimizing the performance when handling large PDF documents. This is solved by breaking it into small chunks using 'Recursive Character Text Splitter'
Accomplishments that we're proud of
I am proud that Pdf Answerer is successfully working as per the Idea I had and I am proud that it can save time and effort of students or professionals.
What we learned
Through this project, I have learned valuable lessons about text processing, AI model integration. This project improved my experience on Python language and Streamlit.
What's next for Pdf Answerer
Pdf Answerer can be updated to support different document formats and languages. It can be developed to support questions and answers in user's preferred language. Or it can be made as applications in different purposes like medical or research contexts.
Log in or sign up for Devpost to join the conversation.