Inspiration

  • Easy, efficient way to organize/store notes especially for those with handwritten notes
  • As students ourselves, possessing the innate need to use better study tools
  • Different modes of studying: Cramming & Comprehensive

What it does

  • A comprehensive website for studying any content the user desires
  • Converts digital blocks of texts to meaningful and concise question/answer pairs
  • Leverages OpenAI API to generate questions based on user input
  • Displays “flashcards” in a pretty format for maximum retention

How we built it

  • We integrated OpenAI / Replicate API into javascript and Python apps and hardcoded a prompt that generated flashcards based on user input. The Python app was hosted on a website with Flask. CSS was utilized to design the front end of our sites.

Challenges we ran into

  • Running out of free OpenAPI tokens / free time on replicate
  • Integrating APIs into our apps, choosing which tools to use (HuggingFace & Django weren't effective)
  • Combining out separate "side-projects" into one comprehensive website, different outputs

Accomplishments that we're proud of

  • Leveraging OpenAI models to generate meaningful questions based on user text.
  • Hosting our app on an actual website rather than having user input data in a terminal.
  • Designing an animated flashcard for our website with CSS

What we learned

  • We learned that APIs are very efficient/useful because they already come with pre-made assets
  • We learned how to host Python scripts onto a website using Flask
  • We improved our communication and teamwork skills by brainstorming ideas and splitting up tasks amongst ourselves
  • We learned more about the inner-workings of AI and how pre-trained AI models (HuggingFace) can be utilized in apps.

What's next for Flaredeck

  • Combine the animated flashcard frontend with our javascript code seamlessly
  • Allow support to turn handwritten text into flashcards (Computer Vision)
  • Creating practice quizzes on top of flashcards
  • Having different modes of studying based on what the user wants (Cram & Comprehensive)

Built With

Share this project:

Updates