GlucoLog

Synapse Health join sleep, exercise, glucose and diet data; Gemini AI spots patterns and suggests actions, with Web3 marketplace lets users spend Health Credits on doctors, trainers and dietitians

Comment

Inspiration The inspiration for Synapse Health (GlucoLog) stems from a deeply personal experience. As a diabetic myself, I understand firsthand the constant need for help and the daily struggle to manage the difficult moments this condition presents. It's an overwhelming reality that involves constant glucose monitoring, meticulous meal planning, exercise regulation, and sleep tracking. Often, this data is recorded in separate apps, notebooks, or simply kept in memory, making it nearly impossible to see the big picture and understand the complex interactions between them.

We realized that the real challenge isn't just collecting data, but finding support and clarity amidst the uncertainty. The inspiration was to create a tool that moves from being a simple, passive data log to a proactive and intelligent health companion and assistant, empowering users to truly understand their bodies, take control of their well-being, and feel they are not alone on this journey.

What it does Synapse Health is an all-in-one diabetes management platform that transforms scattered data into actionable insights. The application focuses on four fundamental pillars:

Unified Logging: Allows users to easily log their glucose levels, meals, physical activity, and sleep hours in a single, intuitive dashboard.

Gemini AI Analysis: This is where the magic lies. We use the Gemini API to analyze the user's data collectively and offer three key high-value features:

Daily Analysis: Provides a summary of how the day's actions (diet, exercise) impacted glucose levels, identifying key connections.

Meal Suggestions: Acts as a personal nutritionist, suggesting healthy and diabetic-friendly recipes based on the ingredients the user has on hand.

Medical Appointment Prep: Helps users formulate intelligent and specific questions for their doctors, based on their own data, to maximize the value of each consultation.

Web3 Marketplace for Specialists: Integrates a decentralized marketplace built on the Polygon network. Users can purchase "Health Credits" to acquire tokenized services directly from certified doctors, personal trainers, and nutritionists, ensuring transparent and secure transactions.

How we built it We built an interactive and functional prototype to demonstrate the core concept, using a modern and accessible tech stack:

Frontend: HTML5, Tailwind CSS, and vanilla JavaScript to create a clean, responsive, and aesthetically pleasing user interface with a "glassmorphism" design.

Artificial Intelligence: The core of our application is powered by the Google Gemini API (specifically gemini-2.0-flash). We carefully designed the prompts for each feature, ensuring the AI's responses were relevant, safe, and presented in a useful and friendly format.

Web3 Infrastructure (Conceptual): The marketplace concept is based on the Polygon network due to its low transaction fees and robust ecosystem. The model involves specialists "minting" their services as tokens (service NFTs) that patients can purchase.

UI/UX: The design focused on clarity and simplicity, aiming to present complex health information and AI insights in a way that is easy to digest and not overwhelming.

Challenges we ran into Prompt Engineering: The biggest challenge was designing prompts for the Gemini API that consistently generated safe, useful, and correctly formatted health advice. This required multiple iterations to instruct the model to act as a "health coach" or "nutritionist" without giving direct, unverified medical advice.

User Experience (UX) with AI: Presenting AI results in a way that inspires trust and is easily actionable was key. We avoided raw data dumps in favor of clear summaries, key takeaways, and lists of recommendations.

Simplifying the Web3 Concept: Integrating the blockchain in a way that felt like a natural enhancement and not an unnecessary complication. The challenge was to explain the value (transparency, security, specialist empowerment) without entangling the user in technical crypto jargon.

Accomplishments that we're proud of Holistic Integration: We are proud to have created a concept that truly unifies the pillars of diabetes management, moving the needle from simple monitoring to comprehensive understanding.

High-Value AI Features: Instead of just showing graphs, we implemented three distinct AI-powered features that solve real problems: understanding the "why" behind the numbers, overcoming "decision fatigue" at mealtimes, and having more productive medical consultations.

Functional Prototype: We built an interactive demo that not only looks good but works. It allows anyone to experience the power of AI analysis with just a few clicks, making the app's value immediately tangible.

Innovative Business Model: The conceptualization of the tokenized marketplace is an accomplishment we are particularly proud of, as it represents a forward-thinking approach to freelance health services.

What we learned The Power of LLMs in Health: We learned that Large Language Models like Gemini are incredibly effective at detecting patterns and correlations in health datasets that a human might easily overlook.

The Responsibility of AI: Working with health data carries an enormous responsibility. We learned the critical importance of setting up safety guardrails in prompts and always positioning the AI as an "assistant" or "coach," not a substitute for professional medical advice.

The Future is Connected: This project reinforced our belief that the future of health applications lies not in isolated tools, but in connected platforms that integrate data, intelligence, and access to human experts.

What's next for GlucoLog The journey for GlucoLog is just beginning. Our next steps focus on taking the prototype to a fully developed platform:

Hardware Integration: Real-time synchronization with popular Continuous Glucose Monitors (CGMs), smartwatches, and fitness trackers via Bluetooth for effortless, automated data collection.

Multimodal Meal Analysis: Implement Gemini's multimodal capabilities to allow users to simply take a picture of their meal. The AI will analyze the image to estimate nutritional information and log the meal automatically.

Full Web3 Marketplace Development: Build and audit the smart contracts on the Polygon network so specialists can mint, list, and sell their tokenized services, and users can securely purchase and redeem them within the app.

Long-Term Trend Analysis: Expand the AI engine to analyze data over weeks and months, offering users insights into long-term trends, progress milestones, and predictive recommendations to prevent future issues.

Built With

  • bolt
  • supabase
  • typequery
Share this project:

Updates