Inspiration: College stress affects pretty much every student in all of America. It's hard to stay organized with your achievements and goals. In addition, many students don't have the financial resources for an expensive college counselor to help them along the way. I took these principles into mind when creating Scholarly.
What it does: Scholarly allows users to organize their achievements, tasks, and future goals, in both academic and extracurricular spaces. They can also view important resources and tools that we give students to increase their knowledge about the admissions process. However, the most important part of Scholarly, is the personalized AI college consultant. With this bot (running on OpenAI's GPT 3.5 Model), students can ask any question they want about college! And the best part - it doesn’t require an expensive college counselor! Whether they are looking for scholarship opportunities, or new extracurricular activities, the AI bot will give them the best response based on their location, major, and interests.
How we built it: Scholarly was designed in Figma, then the UI/UX was created in Flutterflow, and all the custom code and API integrations was completed using Google's Dart language. The API model we used was OpenAI's GPT-3.5 Turbo AI. In addition, I experimented with Lottie Animations for the first time, and integrated it to make the user experience more enjoyable.
Challenges we ran into: The hardest challenge when building this app, was integrating the GPT API and creating the chat system for the bot. For instance, the bot would not take knowledge from previous chats in the same conversation. To fix this, I streamlined the entire prompt into one string straight into the bot. By doing this, I could have the bot act as human-like as possible, and have good flowing conversations.
Accomplishments that we're proud of: I am most proud of making the AI College Consultant. It was a long process, but it now works perfectly smooth. In addition, I was able to fine-tune the bot to only talk about college-related opportunities and not to go off topic. By doing this, the bot feels like a real college consultant!
What we learned: I learned a lot about design and animation when creating Scholarly. I learned about new UI elements in Flutterflow, and how animations work in Lottie. By bringing these elements together, I was able to create an amazing UI that guides the user to whatever they are looking for with ease.
What's next for Scholarly: In the future, I hope to continue to develop Scholarly and add new features. In addition, I want to add authentication so many students can try it out for themselves.
Log in or sign up for Devpost to join the conversation.