Charizzma.tech

Inspiration

In today’s fast-paced world, it is important that we make time to live; and form human connections. However, talking to others can be a challenging or even scary task - especially in new environments, like university!

Because of this, we wanted to create a chaRIZZma trainer that could be used in the comfort of your own home, while letting users practice for real-life social interactions. Be it to make a new friend, deliver a game-changing presentation, or talk to that cute guy/girl, there are so many situations that you can start feeling comfortable with after practicing with Charizzmatech!

What it does

Our personal assistant takes your voice prompt and gives instant feedback to help you elevate your chaRIZZma game. Simply press a button, say your opening line, and receive personalized comments on your prompt!

How we built it

We started off with speech to text using the Speech Recognition library from python. After that we tested out a lot of different models and prompts with Cohere in order to create a good balance in the feedback to the openers.We made our front end using Taipy and added our Speech Recognition and Cohere to it. Lastly we deployed our project to our domain from DotTech Domains with the flask from Taipy.

Challenges we ran into

Due to a lack of documentation and support we weren’t able to do a full deployment of our web app and had to find a work around. We had to locally host our web app and then tunnel it onto another laptop and locally host it there. From this, we were able to reroute our domain to that locally hosted website and use our domain.

We initially had a very different idea that did not work out so pivoting caused us to have a limited time where time management was crucial for us to finish the product on time. In addition, the AI was not good at judging the openers in the beginning and it took a lot of tweaking to the prompts and models to figure out which one would work the best.

Accomplishments that we're proud of

We are very proud of our final product, and especially the work it took to surmount the various challenges we faced while using new technology and methods. Overall, we’re happy with our working, web-deployed demo. This project really pushed our skills and knowledge, and we’re happy with the way it turned out.

What we learned

We learned a lot during this project: not just on the technical side of learning new technologies, but also learning to work with a new team on a tight timeframe. We learnt how to use a new library in Taipy, and workarounds to challenges like deploying the app to the web, which really pushed our coding skills. The tighter timeframe we had really pushed our adaptability and cooperative skills, forcing us to learn how to work more closely and effectively together, turning us into a more close-knit team.

What's next for Charizzma.tech

The journey for Charizzma.tech is far from over. We aim to push the boundaries of simulated human interaction by expanding this into a virtual/augmented reality application to track nonverbal features like eye contact, facial expressions, and body language. This will not only enhance how the user sounds but also how they’re seen, building confidence and improving communication one interaction at a time.