Relay

Log In Screen
Call History
Call Connection
Live Transcription

How we built it

We built Relay by combining speech recognition modules and a backend to control operations with a variety of frameworks. Through extensive testing and prototypes we settled on using OpenAI’s whisper, Kivy, and Django to create an efficient and powerful tool.

Challenges we ran into

Some of the challenges that our team faced was the implementation of whisper and its accuracy of transcription and translation along with its speed. We wanted the transcription and translation service to be as accurate and fast as possible so that when people in a community talk to others, they are able to understand fully with no barriers in existence. Additionally, the implementation of this API and the other packages, along with a call and chat feature inside the app is another challenge itself due to the combination of many different functions being difficult overall.

Accomplishments that we're proud of

Some of the accomplishments that we are proud about was that our group was able to utilize the OpenAI whisper API as well as incorporating the Django library so that the app is able to seamlessly organize data about users. Additionally, we are proud that our transcription service is able to help remove barriers between deaf people and non-deaf people in terms of communication.

What we learned

One of the most important things we learned while working on this project was empathy and the importance of understanding the situations that other people face on a daily basis. After we are able to overcome this initial obstacle, then we are able to truly begin coding with passion, understanding, and a vision for improving communities that we may not even be a part of. Another important element of what we learned are all of the technical and teamwork skills that we acquired along the way, as we worked within the time limit.

What's next for Relay

In the future, our team will expand Relay to the web, and make a website that has the same functionality as the mobile application currently does. We will also expand the application to the IOS platform by adding Cupertino features to make it look native on Apple Devices.