Inspiration
ChatGPT not being able to get the context you need to answer your burning questions about courses.
What it does
Project Alexandria addresses this challenge by converting PDF files into embeddings using OpenAI and then uploading them to Pinecone for effortless retrieval and querying with ChatGPT.
These technologies are integrated with Python Flask, Firebase, and Next.js to add an intuitive UI to interact with the PDFs.
How we built it
AI/ML: Python, Langchain, OpenAI, and Pinecone
Backend: Flask and Firebase
Frontend: Next.js and Tailwind CSS
Challenges we ran into
Navigating Pinecone for the first time presented a learning curve, and working within OpenAI's rate limits posed testing challenges, adding complexity to the development process.
Accomplishments that we're proud of
- Successful conversion of PDFs to embeddings and their efficient storage on Pinecone
- Creation of a visually appealing and user-friendly frontend using Next.js, Tailwind CSS, and Figma
What we learned
- PDF to vector embedding conversion techniques
- Use of Tailwind CSS
- Development with Next.js
- Utilization of Pinecone for efficient data retrieval and storage
What's next for Alexandria
Continued exploration of additional functionalities for PDFs beyond simple querying, such as practice question generation.
Log in or sign up for Devpost to join the conversation.