Inspiration

In our day to day lives we hear an enormous amount of speech; much of that speech is hateful and divisive. It can make people feel marginalized, degraded, or just plain bummed out. This app was designed with the hope that when we see what we actually sound like and get called out for being mean, we can start to build bridges and strengthen our communities.

What it does

When you go to the page, it prompts you for your phone number, and you can hit a button for the app to start listening to you. When you start saying uncool things, you get a text message notifying you that you need to relax, bro. The messages change depending on the harshness of your speech. They also contain the exact words that you said that were uncool, in case you don't remember.

How we built it

Using PyAudio, we stream audio for analysis in real-time. We then use Google-Cloud API's (NLP and Speech to Text) to extract words from the audio and run semantic analysis on it. If the sentiment score of a block of speech is lower than certain thresholds, personalized messages are sent to the user (based on the phone number they inputted) telling them to relax.

Challenges we ran into

Our team didn't have much experience in front-end, so creating a nice, relaxing UI was difficult. We also had trouble getting the audio to stream in real-time for analysis.

Accomplishments that we're proud of

We're very proud of how the website looks and how cleanly the app runs.

What we learned

We learned how to use Google-Cloud API's, Flask, and Bootstrap/CSS. Some team members also learned how to use git.

What's next for relax, bro

To get the text-messages to send to any inputted phone number, we would need to buy a Twilio subscription. We would also want to make it into a mobile app so that it could more seamlessly fulfill its purpose.

Share this project:

Updates