Vitalis Bio — Reversing the Human Aging CurveInspirationWhile chronological age measures the exact rotation of the Earth around the Sun since birth, biological age measures the functional state of human physiology. Modern consumer health is flooded with fragmented biometric streams—Apple HealthKit metrics, continuous glucose monitors, and periodic blood panels—yet users lack a unified, actionable metric that synthesizes this data into a clear longevity vector. Vitalis Bio was born from a simple premise: what if tracking your healthspan felt as intuitive and engaging as managing a financial portfolio? By translating complex clinical biomarker algorithms (such as the Levine PhenoAge model) into a dynamic Biological Age Score, Vitalis Bio empowers individuals to quantify, gamify, and actively reverse their biological aging curve. What It DoesVitalis Bio functions as an autonomous longevity dashboard and lifestyle optimization engine: Passive Biometric Synchronization: Seamlessly ingests continuous physiological signals (resting heart rate, HRV, VO2 max, sleep architecture, and daily steps) via HealthKit and Google Health Connect. Multimodal Lab Parser: Ingests un-standardized lab PDF documents or blood panel reports (HbA1c, hs-CRP, lipid profiles, albumin, alkaline phosphatase) using Gemini vision capabilities to extract structured biomarker values. Biological Clock Engine: Computes a real-time Biological Age Score based on validated clinical aging clocks, comparing biological health directly against chronological baseline. The Longevity Delta Loop: An automated weekly recommendation engine that identifies individual biometric bottlenecks and delivers the top 3 targeted lifestyle, protocol, or supplement interventions to move the needle. Viral Proof Cards: Generates monthly "Wrapped"-style visual summaries designed for easy social sharing. How We Built ItVitalis Bio is engineered with a dark-mode-first, minimalist UX built on a modern full-stack architecture: Frontend: Built with React Native and Tailwind CSS (NativeWind) to deliver a fluid, high-frame-rate mobile UI featuring interactive historical trend graphs (30-day, 90-day, 1-year) and micro-animations. AI & Document Processing: Integrates the Gemini 1.5 Flash API via Google Cloud to execute multimodal parsing of complex, multi-page lab PDFs into validated JSON schemas. Core Algorithm (Clinical Aging Clock): Implements the Levine PhenoAge clinical model. Biological age is calculated by mapping nine physiological biomarkers (albumin, creatinine, glucose, hs-CRP, lymphocyte percentage, mean corpuscular volume, red cell distribution width, alkaline phosphatase, and white blood cell count) into a mortality risk score using a Gompertz mortality distribution. Backend & Storage: Node.js microservices running on Google Cloud Run, utilizing Firestore for persistent document storage and real-time syncing. Challenges We Ran IntoUnstructured Lab Data Variance: Standardizing lab PDFs across hundreds of commercial diagnostic providers (Quest, LabCorp, local clinics) proved difficult due to varying units of measurement and document layouts. We resolved this by implementing grammar-constrained decoding schemas with the Gemini API to enforce strict JSON structure outputs. Biometric Signal Noise: Raw wearable data (HRV and resting heart rate) exhibits high daily volatility. To prevent erratic fluctuations in the daily Biological Age Score, we introduced exponentially weighted moving averages (EMA) across 7-day and 30-day rolling windows. Sub-Cent Latency & Performance: Rendering real-time trend graphs while recalculating complex mathematical models required offloading mathematical calculations to lightweight, local on-device utility modules. Accomplishments We're Proud OfFlawless PDF Parsing: Achieved over 95% accuracy in automatically parsing un-standardized blood panel PDFs into structured clinical inputs without human intervention. Zero-Friction UX: Successfully built a two-tap onboarding flow that connects native HealthKit/Connect feeds and immediately displays a baseline clock visualization comparing chronological vs. biological age. High-Fidelity UI: Created an intuitive dashboard aesthetic combining the clarity of top-tier financial apps with the engagement of modern fitness platforms. What We LearnedBiomarker Synergy: Clinical aging clocks are significantly more sensitive to systemic inflammatory markers (such as hs-CRP) than raw daily activity metrics. Combining active blood panels with passive wearable streams creates a far more resilient predictive model. Actionable Simplicity: Users are overwhelmed by raw medical data; presenting clear, prioritized "Longevity Delta" interventions drives vastly superior protocol adherence. What's Next for Vitalis BioPredictive Intervention Modeling: Incorporating predictive AI agents to simulate how specific protocol changes (e.g., a 10% increase in deep sleep or a 15mg reduction in hs-CRP) will shift the biological age curve 90 days into the future. Ecosystem Integrations: Expanding direct API integrations to continuous glucose monitors (Dexcom, Freestyle Libre) and DNA methylation/epigenetic testing providers.
Built With
- ai-agents
- apple-healthkit
- bioinformatics
- biometrics
- digital-health
- fastapi
- firestore
- gemini-api
- google-cloud
- google-cloud-run
- google-health-connect
- healthkit
- healthtech
- llm
- longevity
- multimodal
- node.js
- pdf-parser
- python
- react-native
- stripe
- tailwind-css
- typescript
- wearables

Log in or sign up for Devpost to join the conversation.