Inspiration
The inspiration behind our project came from some of the classes we have taken at the University of Michigan, where on exams we are allowed to make cheat sheets to use while taking them. This tool would streamline the process and ensure that less time is used up for manual work and more time can be spent actually studying.
What it does
Our hack takes in images of diagrams and tables you want included, and a PDF of your notes, and using Gemini 1.5 Pro, we turn your notes into a clean LaTeX document that incorporates your notes and diagrams and makes it not only machine-readable, but also beautiful and easy to use for studying. The user can then download both the .tex and the .pdf file, so that they can easily modify and view the document.
How we built it
We combined React.JS and Python to create a seamless experience for users. React along with Bootstrap enabled us to build a dynamic, interactive interface for users to upload notes and view the processed LaTeX documents. On the backend, we used Python and Gemini 1.5 Pro API to convert the files into a clean LaTeX document. Integrating the frontend and backend was a challenge, but using Express.js we managed to take the files that the user inputted and pass those into the Python script, which then outputted the .tex and .pdf files for downloading.
Challenges we ran into
The biggest challenge in our project was the integration between the backend (where the files are processed and outputted in Python) and the frontend (where the files are uploaded by the user) We ran into many issues here, as none of us had a lot of prior experience with backend-frontend integration. However, we were able to solve this problem by learning and implementing Express.js, which facilitated the integration smoothly from our React frontend to our Python backend.
Accomplishments that we're proud of
We are proud that we were able to develop using React.JS for the first time. We harnessed the power of creating dynamic and interactive webpages libraries like Bootstrap and MaterialUI. Finally, we spent a lot of time debugging the integration software through Express.js, so we're proud of making that work.
What we learned
We learned how to develop a front-end from scratch and how to integrate a Python backend and a React.JS frontend using Express.js. Additionally, through working as a team, we learned valuable skills in Git version control.
What's next for Noteworthy.ai
The next steps for Noteworthy.ai would be to take in a whole lecture (video & audio), and then transcribing the audio and video into text, charts, and diagrams. We believe this would boost the versatility of our program and reach to more consumers, but user authentication will be a significant challenge. In addition, we would like to introduce more customization features in cheat sheet creation (like number of pages, depth of material), as we believe this could further harness the power of our Gemini API integration.
Log in or sign up for Devpost to join the conversation.