Vital

because seconds matter...

Comment

Inspiration

Medical emergencies are more common than people might think, and knowing how to respond in these situations can make the difference between life and death. In medicine, there is the idea of the “golden hour”, which illustrates how getting help within the first hour of a medical emergency is crucial. Many people who witness these situations want to help, but don’t know how and need help along the way. Due to the disparity, we decided to develop a tool that converts hesitation into action. Vital was created to empower everyday people to respond confidently in emergencies, bridging the gap between panic and professional medical help.​

What it does

Vital is an AI-Powered emergency guidance app for bystanders. The idea is that during an emergency, a user can select what the emergency is and answer questions about the situation in order to gauge what is happening. Based on those answers, we use AI to generate a list to help the person understand what steps to take next. We are not encouraging those to diagnose disease, but to be able to recognize certain signs of distress in order to help someone before emergency professionals can arrive. We also created a regular get information page if people want to learn outside of an emergency situation, as well as a chatbot that is available to ask more questions about specific medical situations.

How we built it

We built this using react-native and the Google Gemini API. We created the front end with react native and then added Gemini calls based on the emergency situation. We have Gemini reading from specific websites based on the situation in order to obtain the most accurate information possible. We also integrated ElevenLabs as part of our backend for the app.

Challenges we ran into

The main challenge we ran into was understanding how to implement the Gemini API in the chatbot and the other features. We also had to make sure that we were balancing AI information with medical safety, as we don’t want people to take everything at face value, but it should be used as a tool that has accurate information that can help. We also ran into issues with backend integration and making sure that there were minimal merge conflicts. ​

Accomplishments that we're proud of

We are proud of getting the chatbot to work and getting our information that is accurate based on the websites. Similarly, we are proud of the fact that we were able successfully integrate gemini API into these features. We are also proud of creating an accessible UI/UX for those using the app with buttons, colors, and fonts that are easy to understand and read. Additionally, we were able to integrate ElevenLabs as part of our backend, which was hard but rewarding.

What we learned

We learned a lot about how integrate gemini API and what those calls require. We also learned about how to incorporate and call certain dependencies within the frameworks. As for the frontend, we also learned a lot more about how to style react native pages and how to make those calls to create a responsive app. Most importantly, we learned how to stick to solving problems. We did our best to push through and come up with solutions, whether that was pivoting on an issue, asking for help, or working with each other to fix the issue.

What's next for Vital

Next steps for Vital include adding more community resources to the app so that people can connect with others and talk about experiences. Make the checklist and chatbots more responsive and accurate so that people can get even more accurate information promptly. We would also like to add some more emergency features, such as scanning images into the bot and assessing vitals on the scene. Having this information before paramedics are able to make it on scene helps give them a better context when they get there and saves time. Additionally, having 911 when called be able to see the answers to the questions that are listed would also be helpful. Adding these features would enable Vital to increase bystander confidence and decrease emergency response time.

Built With

Share this project:

Updates