Symptom Savior

• 

  Symptom Savior - The AI Health Tracking Guardian

• 

  Symptom Savior App Icon

• 

  Detailed App Screen Views

• 

  SymptomSavior.com

Inspiration

I’m terrible at telling my doctor everything—under bright lights and a ticking clock, I panic, recall only the worst aches and pains; I forget about lingering issues and less obvious symptoms. I wanted a companion that captures the forgettable moments as they happen, so no detail slips through the cracks—enter Symptom Savior, your personal health guardian ✨

What it does

• Log symptoms on the fly: Record severity (1–10), duration, location, triggers, and notes—by typing or voice—so you never miss a beat.
  
• AI-powered pattern spotting: Our engine uncovers trends and correlations in your data—think “That rash flares up after I eat nuts.”
  
• Doctor-ready reports: Generate charts, summaries, and PDF exports to breeze through appointments without awkward pauses.
  
• Smart health chats: Chat with our “guardian AI” for instant, context-aware guidance—sometimes you just need a quick fact-check before dialing your clinic and other times you want a complete summary of everything that happened since the last time you spoke with your doctor.

How we built it

• Frontend: React Native & Expo (TypeScript, Expo Router, Lucide icons) for a clean, intuitive UI that could build on mobile devices.
• Backend & data: Supabase PostgreSQL with Row-Level Security for bullet-proof privacy, Supabase Auth (JWT) for login, and a medical RAG “TxAgent” API powering AI responses. Future Integration of Algorand for sharing medical data with doctors securely.
• Voice & AI: ElevenLabs for speech-to-text and text-to-speech, with plans for TavusAI video avatars—hands-free logging and visual check-ins are the future.

Challenges we ran into

• Healthcare compliance: Implementing HIPAA-ready encryption and strict RLS policies to ensure every user’s data stays theirs alone. We are in the process of achieving this milestone.
• Seamless voice input: Balancing background noise, latency, and accuracy so logging by voice feels natural—even when you’re whispering about sensitive symptoms.
  
• Contextual AI: Teaching our RAG system to weave in your medical history, recent logs, and demographics without overwhelming you with jargon or wasting compute.

Accomplishments that we’re proud of

• Delivered a voice-enabled symptom form that logs detailed entries.
• Created AI health consultations with real-time contextual advice and emergency alerts.
  
• Built offline support so you can log symptoms anywhere, even mid-flight.
  
• Achieved cross-platform parity—available on iOS, Android, and Web—with a single codebase.

What we learned

• Strict RLS policies are challenging but essential for real-world health apps.
  
• Voice UX in medical contexts demands extra care for privacy and clarity.
  
• Integrating RAG AI with personal health data requires fine-tuned prompts and safety guardrails.

What’s next for Symptom Savior

• Always: Trying to improve agent response accuracy and time to achieve the most efficient and accurate responses possible.
• Multiple Languages: Will be available down the road with 11 labs integrations.
• Doctor chat integration: Let the AI prep questions or even triage simple reports directly with your clinic. Prescription management and insurance/pharmacy connections are a possible integration down the road.
• Video avatars (TavusAI): A friendly face that reviews your logs and walks you through prep steps.
  
• Predictive insights: Early warnings for flare-ups, based on patterns only our AI can spot.

Bolt.new

• JUDGE ACCESS: username: savevsgames password: 123456
• A New Account: has been set up under those credentials for judges to use to access the app.
  

Built With

• elevenlabs
• expo.io
• rag
• react-native
• supabase
• txagent
• typescript