MinervAI

MinervAI is a study assistant powered by Google Gemini. It transforms dry, text-heavy course modules and reading material into an interactive, personalized learning experience.

Comment

Inspiration

As a team of online OSU post-baccalaureate students, we all resonated on the struggles of being distance learners. However extensive the OSU post-bacc network is, sometimes it is nice to just have someone there with you to help make distance learning more engaging. This is how MinervAI was born.

What it does

MinervAI, named after the Roman goddess of wisdom, is a study assistant powered by Google Gemini. It transforms dry, text-heavy course modules and reading material into an interactive, personalized learning experience. Just paste in your reading material, and MinervAI will generate a summary of its main points, flashcards, and quizzes. It was designed for distance learners like us, and keeps learning engaging, accessible, and much less overwhelming. It is like having a personal tutor, flashcard app, and quiz generator in one – right alongside you, even when you’re alone.

How we built it

Our team leveraged the Google Gemini API, and tuned its model to produce meaningful summaries, quizes, and flashcards for students. Our backend is built with Python and Flask, handling form submissions and API calls to Gemini. Our frontend uses HTML, CSS, and Jinja2 templates to display the AI-generated content dynamically. We focused on creating a streamlined web-app that was responsive, and beginner-friendly.

Challenges we ran into

This was a first-time hackathon for two of our members. Additionally, we had to quickly get comfortable working as a remote team with varying levels of experience. However, we quickly found that our team had great chemistry, and decided on communication, version control, and collaborative coding processes early on.

We also took on the challenge of learning new technologies and frameworks, such as the Gemini API, Flask, and Jinja2. Styling and aligning web elements properly without JavaScript also proved more challenging than expected! But we persevered nonetheless.

Accomplishments that we're proud of

We are extremely proud that we built a functional and cohesive AI-powered tool from complete scratch within 24 hours. All while learning new technologies and collaborating remotely. Every team member contributed meaningful, and we pushed ourselves beyond our comfort zones. We were able to fully integrate Gemini as the backbone of our app, and deliver a complete learning assistant that actually feels useful and relevant to our daily lives as distance learners.

What we learned

We learned how to quickly leverage a new API (Gemini), how to work with Flask and Jinja in a clever way. We learned the importance of prompt engineering to ensure users are provided with the best experience on our app. We also learned new collaboration techniques, and how to build and debug a full-stack app under pressure. Most importantly, we learned that building a fleshed out application within 24 hours is indeed possible and is incredible rewarding! This is why we love coding.

What's next for MinervAI

We plan to turn MinervAI into a browser extension or web companion that can work directly on learning platforms like Canvas. We would also like to add more functionality into the app, like additional games to increase engagement with the material. Ultimately, we want MinervAI to be a go-to support system for asynchronous and in-person learners.

Share this project:

Updates