Inspiration
We all understand how powerful ChatGPT is right now, and we thought it would be really cool to make it available to directly call ChatGPT for help. This not only saves time, it is also more convenient. People do not need to be in front of a computer to access ChatGPT, simply call a number and that is it. This also has the potential for accessibility, people who have disabilities in their eyes might struggle to access ChatGPT through their computer. Now, this will not be an issue.
What it does
An application that allows users to make a phone call to ChatGPT for easier access. Our goal with this project is to make ChatGPT more convenient and accessible. People can access it with just Wifi and a phone number.
How we built it
We use the TWILIO API to set up the call service. The call is connected to our backend code, which uses Flask and Twilio API. The code will receive speech from the user and translate it into text so that ChatGPT can understand it. The code will feed the text to ChatGPT through the OpenAi API. Finally, the result from ChatGPT will be fed back to the user through the call, and the user may choose between continuing the call or hanging up. Meanwhile, all the call history will be recorded and the user may access them through our website using a password generated by our code.
Challenges we ran into
There were a lot of challenges in the front end, believe it or now. Trying to design a good way to represent all the data that we collected from the calls, and connecting them from the backend to the front end. Also, setting up Twilio was kind of a challenge since no one on our team was familiar with anything about call services.
Accomplishments that we're proud of
We finished the majority of our code at a fairly fast speed. We are really proud of this. And this led us to explore more options. In the end, we did implement a lot more features into our project like a login system. Collecting call history, etc.
What we learned
We learned a lot of things. We never knew that services like Twilio existed, and we are genuinely impressed with what it can accomplish. Since we had some free time, we also learned something about lip-syncing with audio and videos using ML algorithms. Unfortunately, we did not implement this as it was way too much to do and we did not have enough time. We went to a lot of workshops. They had some really interesting stuff.
What's next for our group
We will ready up for the next Hackathon, and make sure we can do better.
Log in or sign up for Devpost to join the conversation.