Scribe

Inspiration

One of our group members ran a tutoring company in high school and noticed that his students liked learning best with peer-to-peer collaboration. However, facilitating that connection was very difficult because of the gaps in students' knowledge. We started Scribe to bridge those gaps and motivate students to learn more effectively as a team.

What it does

Scribe is a comprehensive studying tool that allows students to take notes as usual and reap the benefits of automatic and streamlined collaboration, generative AI, and motivational gamified studying. Scribe combines notes from everyone in the class and puts them into a vector server, which allows AI insights to fill in gaps in the student's notes. It also allows them to ask a question about the subject, and the answer will be generated based on the shared notes of the class, which allows for responses tailored exactly for the class content. Finally, Scribe creates flashcards and learning games using generative AI, which helps students to remain engaged in their learning.

How we built it

We used Pinecone as a vector server to store student notes and make it easy to do a similarity search and OpenAI vector embedding and chat completions for our AI insights. We used AWS and Node.js as our backend, which included authentication, S3 storage, and DynamoDB. For front end, we used React and Chakra UI to create a user-friendly and attractive interface for students to take notes, join classes, and view AI insights.

Challenges we ran into

We struggled to integrate our Python backend using an AWS Lambda function.

Accomplishments that we're proud of

We are proud of setting up our vector server and being able to fetch context from it for our AI insights. This elevates our model beyond a simple GPT call and makes it into something special.

What we learned

For many of us, this was our first hackathon, and we learned how exciting and exhausting an opportunity like this can be. We were also exposed to how to create a project from scratch and integrate all parts of it. Finally, we learned about identifying a user base and catering features to potential customers.

What's next for Scribe

Our next steps include a more involved game system, connecting with our initial user base, and using AI to analyze studying patterns and make them more efficient.

Built With

• amazon-dynamodb
• amazon-web-services
• node.js
• openai
• pinecone
• python
• react
• s3