PDF Chatbot With Gemini AI

I’m building a lightweight web app that lets a user upload any PDF, chat about its content, and see answers generated by Gemini. Here’s the workflow I have in mind:

• User uploads a PDF
• Text is extracted with PyPDF2
• Questions are routed to Gemini, constrained to the extracted text
• Answers stream back into a clean chat interface, with full history retained and scroll-able

I’d like the backend in Python using Flask or FastAPI—whichever you feel yields the fastest turnaround. The front-end can be a simple HTML/CSS/JS layout (React isn’t necessary, but feel free if it speeds things up) as long as it stays minimal, mobile-friendly, and intuitive.

Core requirements
• Single-file or drag-and-drop PDF upload
• Robust text extraction via PyPDF2, handling multi-page files
• Gemini API integration with token limits and error handling in place
• Conversation context preserved server-side so follow-up questions stay relevant
• Clear separation of environments/keys and a concise README for local setup

For acceptance, I’ll spin it up locally, upload three test PDFs, and run a short Q&A session. If everything returns accurate, context-aware replies and the UI remains snappy, we’re good.

Let me know your estimated timeline and any clarifying questions—happy to jump on a quick call if needed.