"When you talk, you give yourself away. You reveal your true character in a picture which is more true and realistic than anything an artist can do for you."

Taking inspiration from the great Ralph C. Smedley, we have decided to help people grow into the communicators and leaders that the world needs for the challenges of tomorrow. We aim to do this by leveraging innovative speech-to-text and natural language processing technology to help people become better public speakers.

What it does

Filler words are used by people all the time but ignored by the brains of the person listening. We used the Bing Speech API to record a transcript of the voice and then tried to find the errors in speech. The application’s settings can be changed so that ignores fillers like the human mind is trained to.

How I built it

We used Javascript, HTML, CSS, npm, Microsoft Cognitive Services - Bing Speech API, and TextGears API.

Challenges I ran into

We ran into problems with the REST API as it would only allow 15-second audio inputs and had to re-adapt our architecture to work around this. We made the effort to switch to developing with more robust client libraries instead of relying on a REST API that only supports limited functionality.

Accomplishments that I'm proud of

Some on our team had never used HTML or CSS before, yet we were able to learn quickly and contribute to our final code.

No one on our team had worked with the Microsoft or TextGears API, and we are very proud that we were able to create a working project in just a weekend!

What I learned

Javascript, TextGears, Microsoft Speech APIs, HTML, and CSS and mpm were some of the important new technologies we crash-coursed on. But more importantly, we learned how to work well as a team by learning how to communicate more openly and how to efficiently schedule and plan so that our project would be completed on time.

What's next for Spitch Perfect

Due to time and resource constraints, our first next step will be to improve the user interface further in order to beautify it for our potential users. We also hope to implement a VR element to the product to simulate a more stressful environment for the user. With increased stress, the user would be more error-prone and we want this so that the user can recognize and eliminate these errors even in the most challenging environment.

We hope to have a sustainable business so users can continue to enjoy the product and so we can continue to improve it. Our business model will be a freemium model in order to encourage new users to experience the product. The premium version of the product will have added functionality such as custom text analysis that lets them choose certain words they wish to filter out. We will also limit the amount of times users can use our product. By upgrading it to premium, they will have unlimited uses.

Another source of revenue would be advertisements from vocal and presentation coaches who are looking for new customers.

Built With

Share this project: