PDFinance

• 

  Logo

• 

  Upload Screen

• 

  Analysis Screen

PDFinance

Inspiration

• Financial analysts, investors, and professionals deal with massive amounts of documents (SEC filings, earnings reports, contracts).
• Searching for specific insights manually is time-consuming and inefficient.
• We wanted to build a smart, AI-powered tool that makes finding answers to document-based queries fast, accurate, and effortless.

What it does

✅ Uploads & stores PDFs in an SQLite database (financial reports, contracts, whitepapers) ✅ Extracts and processes text from pages based on a user's query ✅ Uses vector-based search to find the most relevant documents ✅ Generates intelligent answers from the retrieved content using an AI model ✅ Provides an intuitive Tkinter GUI for seamless user interaction

How we built it

💾 Database: SQLite3 for storing PDFs and extracted text 📄 Text Extraction: PyPDF2 for parsing and extracting PDFs 🧠 Vector Search: LangChain for document processing and AI interactions 🤖 AI Model: OpenAI’s API (ChatGPT) for generating responses based on retrieved information 🖥️ Frontend: Tkinter for an interactive user interface 📂 File Management: Shutil for file operations 🔑 Environment Variables: Dotenv for managing environment variables

Challenges we ran into

🚧 Integrating multiple components 📉 Working with objects to analyze vectors 🔄 Refactoring & redoing multiple sections to increase efficiency

Accomplishments that we're proud of

🏆 Successfully built a working AI-powered document search tool! 📊 Achieved accurate retrieval of insights from PDFs 💡 Developed a scalable solution that could be extended to other industries 🚀 Learned how to integrate vector search & AI models effectively

What we learned

📌 Implementing a local SQLite3 database and PyQt6 GUI 📌 Integrating vector search with AI models 📌 Optimizing systems for real-time performance

Built With

• chatgpt
• dotenv
• functool
• langchain
• openai
• pycharm
• pypdf2
• python
• shutil
• sqlite
• tkinter