Speakeasy

This was our basic thought process on how we wanted it to be designed.

Inspiration

As students, we are constantly needing to attend interviews whether for jobs or internships, present in class, and speak in front of large groups of people. We understand the struggle of feeling anxious and unprepared, and wished to create an application that would assist us in knowing if we are improving and ready to present.

What it does

Our application is built around tracking applications such as body language, eye movement, tone of voice, and intent of speech. With the use of analyzers, Speakeasy summarizes and compiles results along with feedback on how to improve for your next presentation. It also grants the ability to learn how your presentations come across to others. You can see how clear and concise your speech is, edit your speech's grammar and spelling on the fly, and even generate a whole new speech if your intended audience changes.

How we built it

We used LLP models from Cohere to help recognize speech patterns. We used prompt engineering to manufacture summary statements, grammar checks, and changes in tone to deliver different intents. The front end was built using React and Chakra UI, while the back-end was created with Python. The UI was designed using Figma and transferred on to Chakra UI.

Challenges we ran into

There were numerous challenges we ran into in the creation of this application. With it we learned to quickly adapt and adjust our plans. Originally we planned to utilize React Native, however, were faced with installation issues, OS complications, and connectivity between devices. We then pivoted to Flutter, however we encountered many bugs within an open source package. Finally, we were able to move to React where we successfully created our application. For the backend, one of the biggest challenges was guessing the appropriate prompts to generate the results we need.

Accomplishments that we're proud of

We are proud of our ability to adapt and react to a multitude of difficulties. As half of our team are first-time hackers, we are proud to have been able to produce a product that our whole team is satisfied with. This type of idea-making, app development process, and final product creation are experiences we had never encountered beforehand.

What we learned

We learned how to think creatively in difficult and time-crucial situations. As well, how to design using Figma and Chakra UI. We also learned from each other as teammates on how to collaborate, generate ideas, and develop a product we are proud of.

What's next for Speakeasy

We hope to both add on to and improve our features as well as making it accessible for people around the world who speak different languages. We look forward to continuing to develop our tone analyzers and speech intent translators to increase accuracy. Along with strengthening our body language trackers and connecting it with databases to better analyze the meaning behind movements and how it conveys across audiences.