-
-
Landing Page
-
Upload Notes
-
Home
-
Flashcards
-
Character Select
-
Study Group
Inspiration
Why in the 21st century do we have to still use flashcards in the same old way? Traditional flashcards demand manual input that forces you into writing them down yourself and digital websites such as Quizlet bury AI features hidden away through layers of buttons and advertisements.
Introducing Study Rush, the modern, fully AI-integrated flashcard application. With our AI solutions, we transform the tedious process of creating study materials into an effortless, intuitive experience. After just a few uses, you’ll spend less time rewriting your notes and more time actively practicing your material. Let’s make study smarter.
What it does
Study Rush is a study partner at your fingertips. Start at the landing page and login to save your flashcards and get started on our website. Our product presents three new AI solutions:
Automatic Flashcard Generation
Copy and paste your notes into a text box and let our AI make flashcards for you instantly. It takes your text to the backend where our Large Language Model (LLM) implementation will convert it to flashcard format and save it to a database.
Quiz Mode
In addition to flipping cards back and forth, Quiz mode adds another dimension to flashcard studying. Test your knowledge on a concept by typing into a textbox and get immediate feedback from our AI.
AI Study Groups
We introduce a novel concept with chat conversations. Nearly every chat bot application will have a 1-on-1 assistant. But for study groups, you want multiple opinions to go off of, listen to others speaking, and help others as much as getting help yourself. Study Rush gives you the freedom to organize a session with AI study partners with distinct personalities who don’t always know everything about your flashcards.
You can select which two characters you want to form a group with to discuss your flashcard deck. Ask questions and give answers just like a real group. The fun and variety of conversation between a group will help you remember topics better.
How we built it
We started with brainstorming ideas for the hackathon and decided on Study Rush. In pre-planning, we made a basic UI wireframe (a quick drawing), determined tools we wanted to use, and organized time to complete the features we planned between only two group members.
During the hackathon, we continued to plan the UI, actually made it in React, developed the frontend and backend of the website, tested the AI features, and applied our own notes into Study Rush.
Frontend: node.js, React, Tailwind CSS
Backend: Python, Firebase Authentication, Flask for API, SQLAlchemy, PostgreSQL, Neon (DB hosting), Anthropic Claude 3.5 Sonnet
Other: GitHub for version control, Docker and Back4app for deployment and hosting, Postman to test backend API
Challenges we ran into
Although we were able to complete our MVP, it was not smooth sailing as we ran into different issues that slowed our development. One issue that still persists throughout the project is difficulty with the setting up a consistent UI. Our UI was difficult to configure with different pages having slight differences in similar formats. The time constraint also led to difficulty as we did not think we could train or fine tune models as we wanted to thoroughly build our application.
Accomplishments that we're proud of
We finished our MVP in the allocated amount of time. Study Rush has a clean UI which seamlessly connects with our server and postgres database. The server was thoroughly tested with unit tests and the UI was carefully crafted in Figma. We are proud of the preparation and work done during the development of Study Rush.
What we learned
We learned how to work as a team to develop a product quickly and thoroughly in a short time span using both known and unknown technologies.
What's next for Study Rush
Study Rush is an ambitious app that specifically focuses on the summarization task. There were a few fundamental features that we wanted to add but ran out of time for: editing flashcards, deleting decks, deck collaboration, deck sharing, more unique characters in AI study group, custom characters, mobile compatibility, and so on.
We’d also like to improve the UI and make it more active in the future to make it more satisfying to use and study with. The overall application should have user feedback and testing to ensure that it properly helps people who want to use our service.
We could enhance each card in Quiz Mode with statistics such as percent accuracy and confidence level from text, as well as add a calendar for planning and tracking daily practice streaks. Additionally, incorporating quick feedback forms would help us determine if the flashcards and conversation AI are effective or need improvement. Other features to implement include a timer to introduce a timed activity element to practice and a LATEX render module for viewing latex equations on the flashcards.
In terms of security, we need to ensure the API for registering, login, and deck content is secure. Privacy features for keeping decks hidden and preventing abuse on our LLM API usage are important to keep our application running. We also need to be responsible with our AI usage: preventing prompt injection, adding disclaimer tags when you use our AI products, and making sure that users under 18 who use our products are kept safe.
Further improvements to our app could be that we gather training dataset and fine-tune a language model for the specific educational tasks that students need. We could also implement a RAG system where the model is able to reference the exact chunk it is citing.
Log in or sign up for Devpost to join the conversation.