SymptomSnap — AI Health Symptom Checker

• 

  App Logo

Inspiration

A lot of users are now using smartphones. Most times people ignore early symptoms attributing them to either fatigue or just a normal occurrence

What it does

Through use of AI, users describe symptoms via voice or text; AI suggests possible causes, severity levels, and next steps. Includes a body-map for pointing to pain areas.

Symptom intake — chat-style text intake with quick-answer chips Voice input — on-device speech-to-text (iOS Speech / Android SpeechRecognizer); audio never leaves the device Body map — tappable front-view body silhouette with stable semantic region IDs mapped to standardized codes Health profile — age, sex, conditions, medications stored locally with validation AI triage pipeline (in development) — pure-LLM analysis on Hugging Face Inference Providers with structured JSON output, backed by a deterministic red-flag rules engine that always overrides the LLM toward emergency care

How we built it

I will build it solo using React Native. I will connect relevant LLM models to answer user queries. I also used OpenCode Agentic tool for coding

Challenges we ran into

For Auth, this was the first time to use Supabase with React Native. Supabase had revamped their platform and that caused some problems for me.

Accomplishments that we're proud of

I tested it in a local dispensary. The nurses have adopted it as part of their work

What we learned

You need to verify and settle on the right framework. I used Metro instead of Expo and experienced loading speeds

What's next for SymptomSnap — AI Health Symptom Checker

Built With

• llm
• node.js
• opencode
• react