Inspiration
With hours and hours spent digging through course content before exams, we find that one of the biggest struggles for students is finding an efficient way to summarize course content in a comprehensive and concise way. We created hop-to-it with this issue in mind, hoping to make it a perfect study tool to prepare for exams.
What it does
hop-to-it is a study tool that facilitates students with their note-taking process by allowing them to upload their notes or lectures in various formats and convert their content into a summarized form for easy comprehension. It accepts three different forms of media as input: audio (ideal for transcribing and summarizing records lectures), images (for digital and written notes), and text (for summarizing large text documents). After a file is uploaded, the contents of the files are transcribed and displayed. It then generates a summary of the captured text or speech, ensuring that all its key points are preserved.
How we built it
- React/Javascript Frontend with the Material UI library and Flask/Python for the backend server for processing requests
- OpenAI's API for text summarization
- Speech Recognition package for speech-to-text analysis in Python
- Python-tesseract OCR tool for recognizing text in images
Challenges we ran into
One of the main challenges we ran into was connecting the frontend to the backend. We wanted to direct our focus to our backend and made a simple, temporary UI with HTML to test their functionalities at first. We then made the shift to react and had difficulty connecting our new frontend to our Flask server. Because of this, we encountered errors processing our requests and spent a lot of time debugging and resolving these errors.
Accomplishments that we're proud of
With this being the first time working with Flask and React, we were proud of developing a functioning web app with much prior knowledge. With many of our past projects being frontend heavy, we were also glad to have made a app with more backend functionalities
What we learned
One of the main skills that we learned from Hack Princeton was utilizing flask as the backend server and linking it to a frontend that is built using React and Javascript.
What's next for hop-to-it
Some functionalities for future development that we can add to the program would be allowing users to store 'favorite' summaries and view their upload/generate history in the form of a database. Moreover, we also discussed the possibilities of publishing hop-to-it as an online platform that allow users to share their notes to other users for the purpose of collaboration and studying. Another improvement that can be done is to develop a more consistent output from the AI. After testing out the OpenAi api, we found that the text generator used to summarize the paragraphs may be inconsistent and generated inaccurate output from time to time. Therefore, we could look more into the model or can improve the program by trying to build and train a model, in order to achieve a better and more consistent output from AI generations.
Log in or sign up for Devpost to join the conversation.