Prometheus

Welcome to Prometheus—an AI-powered study app that analyzes videos, generates customizable quizzes, and tracks your progress across sessions to make studying smarter, faster, and more efficient.

Comment

Inspiration

As STEM students, every member of our team has heavy coursework and has difficulty finding a school-life balance. Additionally, many of us have had professors that did not lecture well and resorted to online lecture videos. To make this kind of self-learning more feasible, we made Prometheus: an app that extracts information from videos and can make quizzes for spaced, repeated testing. Prometheus saves time spent watching (and re-watching) long lectures, and improves knowledge retention by providing spaced repetition.

What it does

Simply upload your documents or videos to Prometheus. Our AI app will analyze your content, extracts key points, and generates concise summaries. Then, it crafts quizzes tailored to your difficulty preference, helping you reinforce your understanding and improve retention.

How we built it

  • Frontend: Streamlit
  • Backend: Python
  • Database: Firebase
  • AI: Gemini and SambaNova APIs

Challenges we ran into

We encountered challenges with Streamlit's multi-page limitations and the steep learning curve of SambaNova. Additionally, integrating custom CSS and components proved to be a significant hurdle. However, we persevered and learned valuable lessons about AI integration and efficient development.

Accomplishments that we're proud of

We're proud to have successfully integrated SambaNova, a cutting-edge AI API, into Prometheus. This integration has significantly boosted the speed and efficiency of our AI-powered summaries and quizzes, providing a seamless and timely learning experience for our users.

What we learned

We learned to craft precise prompts to get desired outputs from AI models, as well as greatly improving our ability to build functional web applications using Streamlit. We learnt to use SambaNova API for fast integration (and realized very late that we could have just used JSON mode in SambaNova ://)

What's next for Prometheus

Authentication: We'd like to track a user's individual progress over time and ensure data privacy and security through robust authentication and authorization mechanisms. SambaNova: Integrate the JSON mode and replace the Gemini model to make it faster Gamification: We'd like to add more stats, a point system, leaderboards, and possibly badges to encourage learning through gamification and make the application a standalone learning platform that can quickly turn vast amounts of video content into learning material.

Fun Fact:

How we decided on the name Prometheus: because SambaNova makes it blazing fast, we use 🔥(base), and we brought the gift of automated quizzes to mankind.

Built With

Share this project:

Updates