CoreCare

CoreCare - a health website application that understands the intersection of what you can eat and how you feel and turns confusing health conditions into simple and personalized guidance.

Comment

Inspiration

Our team was inspired by the often overwhelming and confusing journey individuals face when managing chronic health conditions like PCOS, IBS, anemia, or Celiac disease. We noticed that generic health apps often fail to account for the complex intersection of dietary restrictions (like gluten or lactose intolerance) and specific medical symptoms. We wanted to create an accessible and supportive companion that simplifies nutrition and health management through the thoughtful application of AI.

What it does

Our platform is a comprehensive health and diet management tool tailored to a user's specific medical profile.

  • Personalized Health Dashboard: Users can toggle specific diagnoses, symptoms, and dietary restrictions to customize their experience.
  • AI Health Companion: A persistent, context-aware chatbot powered by Google Gemini that provides natural remedies and real-time recipe generation.
  • Smart Recipe Forum: Automatically generates three specialized recipes for every health issue a user selects. Each recipe includes metric/customary toggles, detailed prep times, and descriptions of how the meal specifically helps their symptoms.
  • Educational Health Hub: Provides interactive cards detailing various conditions, symptoms to watch for, and "what to eat vs. what to avoid."
  • Crisis Safety: Features hard-coded triggers for self-harm or medical emergencies to direct users to professional lifelines immediately.

How we built it

We focused on creating a seamless, unified experience across the stack:

  • Frontend: Built with React, TypeScript, and CSS, using a calming, earth-toned color palette (#99a98f, #dbead2, #fff8de) to ensure the UI felt inviting rather than clinical.
  • Backend: Built with Express.js and Node.js.
  • Database: MongoDB was implemented to store user profiles, saved health data, and archived chat histories.
  • AI Integration: Integrated the Google Gemini API to power the intelligent chatbot and the dynamic recipe generation engine.
  • Speech Capabilities: Implemented Web Speech API for a mic-enabled "hands-free" chat experience, complete with pulsing visual indicators.

Challenges we ran into

One of the primary hurdles was ensuring the AI provided helpful health advice without crossing into "diagnosis" territory. We had to carefully engineer prompts to ensure the bot suggests natural remedies or OTC options while always prioritizing safety. Technically, syncing the frontend recipe cards with the specific JSON format required by our Gemini prompts—specifically handling the dual metric/customary ingredient arrays—required rigorous testing to ensure a smooth UI/UX. Another challenge we faced was trying to find a health API that was verified by the FDA to provide credible information about dietary instructions and health conditions. So, in order to still incorporate information about dietary instructions and diagnoses for educational and informational purposes, we made our own .json file, using credible data from mayoclinic.org, and implemented attributes such as symptoms, what to avoid, and substitutes.

Accomplishments that we're proud of

  • Intuitive UI Integration: We successfully created a persistent chatbot that lives across all pages without being intrusive, maintaining its state and memory.
  • Dynamic Content Generation: Our system doesn't just show static data; it generates a custom "Recipe Forum" on the fly based on a user’s unique combination of 10+ health conditions and 5+ dietary restrictions.
  • Accessibility Features: Adding speech-to-text and clear, simple navigation makes the app more usable for individuals who might be struggling with symptoms like headaches or fatigue.

What we learned

We gained significant experience in Prompt Engineering—learning how to force an LLM to return strictly formatted JSON for our recipe cards. We also learned a great deal about UI/UX for HealthTech, specifically how to present dense medical information in a way that is readable, concise, and non-threatening. From a team perspective, we learned how to effectively bridge the gap between AI-generated data and structured React components.

What's next for project

  • User Accounts: Finalizing the login/sign-up flow to allow users to access their saved recipes and chat history across any device.
  • Community Forum: Expanding the "Recipe Forum" to allow users to post their own tried-and-tested remedies and rate AI-generated suggestions.
  • Wearable Integration: Connecting the platform to wearable devices to pull real-time data like activity levels or heart rate to further refine dietary recommendations.
  • Expanded Database: Adding more niche health conditions and a wider library of "Health Tab" educational content.
Share this project:

Updates