Inspiration
The inspiration comes from the universal figure of the "Doctor Grandma"—the person who always knew exactly which tea would settle a stomach or which ritual would cure a headache. We wanted to digitize this warmth and ancestral knowledge, combining it with the precision of modern AI to provide a wellness tool that feels more human than clinical.
What it does
The application acts as a personal holistic consultant. Users input their current symptoms or feelings in natural language, and the AI generates a "Care Plan" that includes:
Identification: Possible root causes described in supportive, non-alarmist language.
The Kitchen Archive: A curated list of healing foods to prioritize and triggers to avoid.
Daily Rituals: Holistic lifestyle actions, such as specific rest patterns or breathing exercises, to harmonize the body.
The Archives: A session-based history system to track and revisit previous consultations.
How we built it
We utilized a lean, high-performance tech stack to ensure the app is portable and fast:
Logic: JavaScript handles the application state and the integration with the Gemini 2.5 Flash API.
Styling: Tailwind CSS was used to create a custom "Traditional" design system, utilizing utility classes for rapid, responsive UI development.
Data Flow: We implemented a structured JSON response schema to ensure the AI's output perfectly populates our custom dashboard components.
Reliability: Integrated an exponential backoff algorithm to ensure the application remains stable even during high API traffic.
Challenges we ran into
Tuning the "Voice": It was challenging to prompt the AI to stay within the "Doctor Grandma" persona—being helpful and traditional without making dangerous medical claims.
UI/UX Balance: Balancing a "traditional/old" look with "modern/functional" usability required several iterations of the design system, especially regarding the sidebar "Archives" logic.
API Constraints: Managing potential rate limits required building a robust retry mechanism to ensure users didn't lose their analysis results.
Accomplishments that we're proud of
The Aesthetic: Successfully creating a web app that feels like a physical apothecary shop through CSS and typography alone.
Session Persistence: Building a fully functional "Previous Records" system that allows for instant restoration of complex AI data without a backend database.
Prompt Engineering: Achieving a highly structured and reliable JSON output from the Gemini model.
What we learned
We learned that the "User Interface of the future" doesn't have to look futuristic. By applying modern AI to a traditional aesthetic, we discovered that users often feel more comfortable sharing health concerns with a tool that feels warm and familiar rather than cold and clinical. We also deepened our knowledge of responsive state management in vanilla JS environments.
What's next for Doctor Grandma Guide
Multilingual Wisdom: Translating the "Grandma" persona into multiple languages and cultural wellness contexts (e.g., Ayurveda, TCM).
Photo Recognition: Allowing users to upload a photo of their tongue or skin symptoms for more detailed botanical advice.
Community Recipes: A feature to share "Grandma's actual recipes" based on the AI's suggested ingredients.
Device Integration: Syncing "Daily Rituals" with calendar apps to provide reminders for wellness activities.
Log in or sign up for Devpost to join the conversation.