Noteably

Noteably's Logo
The concept for the future of the user interface for Noteably

Inspiration

Leading up to HackKU 2023, our team had discussed several options for exciting and innovative programs that we could reasonably develop in 36 hours, given that this was all of our first Hackathons. We knew it had to be something that we could find useful, but wanted to keep our options open to match the Theme track. Noteably was born from that idea, it had to be something useful, but we also wanted to compete in the Theme track, so what was an aspect of our lives that we wanted to make more "accessible"? Well we thought the note-taking and learning process (which we do daily as college students), was a great place that could be expanded with the recent development in generative AI.

What it does

Noteably takes in audio files and outputs a condensed note sheet that highlights the key points and allows those who are unable to take notes or those who find taking notes to be a chore to still have the ability to look back on what they learned in various classes, meetings, or interviews.

How we built it

Building Noteably turned out to be a more complicated task then first expected; however, our strategic assignment of tasks and problems to various members of the team helped each of us use our strengths to create the best possible product given our lack of experience in hackathons. However, using this segmentation approach we were able to quickly solve problems and each team member became an expert in different various areas of the program including things like Google Cloud APIs, OpenAI APIs, more advanced I/O, and interface development.

The final product was a combination of Google Cloud Speech-To-Text, being fed into a Generative AI model to cut down the bulk, which was managed by a super intuitive and accessible interface.

Challenges we ran into

There were a lot of challenges we faced when developing Noteably, including a major shift midway through development when the locally ran generative AI model turned out to not be smart enough for what we wanted to do. This led us to shift development towards a more cloud-based approach using OpenAI's APIs. This ended up being good thing in the end, as it means that almost all of the data processing is done in cloud, meaning hardware power is not a limiting factor in using the software, increasing accessibility and overall increasing the ability for people to learn new information, which was the main purpose of this software. Other issues we ran into included unfamiliar types from services like Google Cloud API, which served to make the program way more accessible, at the cost of development complexity working with such a complicated system.

Accomplishments that we're proud of

In the end, we are all very proud of what we created. We not only got to develop the bare minimum we hoped to achieve to show off an awesome product, but also got to expand upon that bare minimum and introduce even more accessibility features including a very simple way to interact with the program.

What we learned

In the end, each person walked away learning different things, but what we collectively learned is the power of working together and delegating tasks to improve efficiency and create the best program possible. Personally, each of us learned a lot, especially when it comes to integrating various systems together and interacting with APIs, it also gave all of us a true look at what software development looks like on a bigger scale then classroom experiences.

What's next for Noteably

Noteably was conceived as something that we actually wanted to use, so we plan to continue developing it and implementing more features to create an awesome user experience. We've looked at how this could evolve into something more than just a fun project, and more into something that you would develop for a major tech company. This includes using the generation time to display ads covering the financial costs of the API calls we are making, and the development of new features, with the biggest being a mobile app putting Noteably in the pockets of billions of people achieving our goal of making education and learning more accessible for all.

Built With

Submitted to

HackKU 2023
- Winner Best Use of Google Cloud

Created by

I worked on integrating all of the different parts of the system to work together. In addition, I wrote the code the managed interacting with the OpenAI API which simplified the text.

Michael S
I primarily worked with integrating the google cloud services with the rest of our program. For this project we need to use both the google cloud storage API and the google speech to test transcription API. My job for this project was mostly to accept an audio file and turn it into text.

Holden Vail
I wrote the pdf_output.py file which takes the notes given from the generative ai model and creates a well formatted pdf file that can be used as notes for the original audio. I also helped organize our presentation for the demo.

Blake Carlson
I worked on the graphical user interface, the early structure of the code, and the early methods that set the foundation for the rest of the project.

Logan Smith