SatChat | Devpost

Inspiration

We wanted to increase the accessibility of LLM technology when people don't have access to the internet.

What it does

Imagine you are stranded or lost without wifi or cellular service and you need to answers to your questions. For example, let's say that you are camping in the woods and forgot how to light a fire. You can't connect to the internet or any AI but, you can use your phone that has satellite connection to send messages to our number which will return answers that Gemini creates. Imagine you are out in the field studying different plants in the amazon rainforest. SatChat has your back, with just a satellite connection you can learn more about specific plants or fauna.

How we built it

We created a google extension that allows us to scrape google voice to get incoming messages from users. By doing this we were able to complete the project without using a SMS API simulator like Twilio which is something we initially had trouble with. When the google extension finds a new incoming message, it sends it to our Flask server which then sends the message to Gemini which returns a response back to our Flask server. We then use Google Voice to send that response back to the user automatically. You can now talk to a chatbot straight through your messages/texts.

Challenges we ran into

While using Twilio, we ran into a lot of issues with setting up the phone numbers to make them work. There were many different permissions required that needed multiple days to get approved. We didn't have time to wait, so we had to pivot and change how we built SatChat. Google voice doesn't have an API, so we had to manually web scrape the page to get the content we needed. We also ran into some issues setting up a Google Chrome extension to work with our backend. It took many revisions to get it to work properly.

Accomplishments that we're proud of

We are proud that we created a useful hack that can help people in situations where they need information quickly and don't have access to great data or wifi. We are also proud that even though we hit a bottleneck that completely halted our progress, we were able to pivot and think of a way to finish building our project without using the "easier" method.

What we learned

Before we even started working on the project, we started to look at our own lives and aimed to solve an issue that people face. With that, we learned to keep the users needs at the top priority throughout the development process. We also learned that it's important to be able to adapt and to keep momentum when working on a project.

What's next for SatChat

Our idea is meant to be simple, so the first thing we would like to do is move from the web extension to a real SMS API provider like Twilio. After that we would love to go public and provide this app for real life users that face the same problems.