Inspiration
Public speaking has always been a skill that we've tried to improve throughout the years. We thought it would be a great idea to create an application to help people practice for their big interview, presentation, or pitch.
What it does
Parlé takes in an audio recording of the user during their practice runs and outputs a scoring system based on the number of filler words used throughout the presentation. Parlé then suggests possible adjustments and fixes of the transcript to improve the dialogue.
How we built it
We used React to build the frontend, Google Cloud Speech-to-text API, and Co:here's natural language processing API to receive the adjusted and improve the dialogue.
Challenges we ran into
Many blockers dealing with Google Cloud Speech-to-text as it was not deploying properly. We also had issues converting an mp3 to a .raw file.
Accomplishments that we're proud of
A working application! Although the backend requires more work, the UI is what we wanted.
What we learned
A lot
What's next for parlé
A lot
Built With
- co:here
- express.js
- google-speech-to-text
- node.js
- react
- typescript
Log in or sign up for Devpost to join the conversation.