FinalYearKit
AI / ML

Chat with PDF

Ask questions about any PDF, get answers cited by page number.

Streamlitpymupdf4llmsentence-transformersFAISSGroq API
chat.py
What does section 3.2 say about the results?
Section 3.2 reports a 14% improvement in accuracy after the retrieval step was added, attributed mainly to reduced ambiguity in multi-part questions.
Source: page 12
TESTED

Screenshots

Chat with PDF screenshot 1
Chat with PDF screenshot 2
Chat with PDF screenshot 3

What it does

A retrieval-augmented chat system: upload a PDF, ask questions in plain English, and get answers grounded in the actual document, with every answer citing the exact page it came from. Built with hybrid retrieval (exact-match + semantic search) so it handles both specific lookups and broad summary questions correctly.

Features

  • Upload any PDF and start asking questions immediately
  • Answers cited by exact page number
  • Hybrid retrieval — correctly handles specific numeric/identifier questions, not just fuzzy topic matches
  • Handles both narrow lookups and whole-document summary questions

What's included

  • Full working Streamlit application
  • 8-chapter Word report (architecture, requirements, testing, references)
  • 13-14 slide presentation deck
  • Viva question bank with cheat sheet and pitch script
  • Architecture diagrams

Pricing

Same pricing tiers across every project kit.

Code Only

₹1,500

The working application, source code, and setup instructions.

  • Full source code
  • Setup/run instructions
  • requirements.txt
Most popular

Code + Documentation

₹2,500

Everything in Code Only, plus the full submission report and slide deck.

  • Everything in Code Only
  • 8-chapter Word report
  • Presentation deck
  • Architecture diagrams

Full Package

₹3,500

Everything you need to submit and defend the project, viva-ready.

  • Everything in Code + Documentation
  • Viva question bank + cheat sheet
  • Customized to your name/college
  • WhatsApp support until submission

Questions about this project

Does it work with scanned PDFs?+

It works best with text-based PDFs. Scanned/image-only PDFs need OCR, which isn't included by default but can be added.

How long does setup take?+

Under 10 minutes with the included instructions — it's a standard Python + pip install.

Message us