SpoonShare

• 

  SpoonShare

Inspiration

People living with rare and chronic illnesses often describe life in “spoons”: finite, unpredictable units of energy. The problem is that this burden is mostly invisible—and hard to communicate to caregivers and clinicians in a way that feels objective. We built SpoonShare to turn daily symptom burden into evidence-based pacing guidance and a shareable clinical narrative.

SpoonShare started as a way to turn “I feel worse today” into evidence: weather shifts, pain, sleep, activity load, and upcoming schedule demands translated into a practical daily pacing plan.

What it does

SpoonShare is an AI-powered energy management platform that helps users: -- estimate a daily SpoonScore from sleep, pain, weather, and profile factors, -- predict crash risk from upcoming events, -- receive pacing guidance before energy collapse, -- share live status with caregivers, -- generate a clinician-ready weekly report. Our core idea: make invisible effort visible, trackable, and shareable.

How we built it

We built SpoonShare using Next.js, React, TypeScript, Supabase, and Groq-powered AI pipelines through LangChain.

A core differentiator is our Symptom-Weighted Engine. To power it, we developed a custom Python ingestion script to scrape the Genetic and Rare Diseases (GARD) Information Center, creating a high-fidelity dataset of rare disease names, symptom descriptions, and validated medical URLs. During onboarding, when a user identifies their condition, SpoonShare does not rely on a static lookup. Instead, it uses Groq’s LPU inference for real-time linguistic analysis of the disease’s symptom cluster. By evaluating symptom intensity and physiological categories, the model computes a personalized Metabolic Tax that adjusts starting SpoonScore. This grounds each baseline in disease complexity rather than one-size-fits-all tracking.

We then combine: -- Morning pre-flight signals (sleep/pain/weather), -- Calendar demand forecasting, -- Dynamic spoon cost modeling, -- Caregiver task-claim recovery, -- Weekly report generation and sharing.

Challenges we ran into

Keeping spoon math and event accounting consistent across patient and caregiver views. Handling missing integrations gracefully (calendar/weather/API keys) without breaking UX. Designing secure share-link workflows for caregivers while preserving privacy. Making AI output structured, reliable, and explainable enough for health-adjacent use cases. Balancing clinical tone with empathetic UX for users managing fluctuating symptoms.

Accomplishments that we're proud of

Built an end-to-end rare-disease-aware pacing product in hackathon time. Created a condition-specific onboarding engine using real clinical symptom context. Delivered caregiver collaboration and clinician-facing reporting in one experience. Converted invisible daily burden into actionable, communicable insights.

What we learned

In health tools, trust depends on consistency and transparency as much as model quality. Personalized baselines matter more than averages for chronic/rare illness workflows. The best AI experiences combine probabilistic insight with deterministic fallbacks. Accessibility and emotional framing are product requirements, not polish.

What's next for SpoonShare

Add confidence scoring and source attribution for each forecast. Expand wearable and passive signal integrations. Improve longitudinal trend detection and intervention suggestions. Run pilot feedback loops with patients, caregivers, and clinicians. Evolve from decision support toward proactive, personalized care navigation.

Built With

• groq
• langchain
• meteo
• next.js
• python
• react
• supabase
• tailwindcss
• typescript
• vercel
• zod