Clover

symptoms log
clinical report
find the care you need
slides

Inspiration

The average Canadian woman waits 2 years to get a PCOS diagnosis. Not because the symptoms aren't there — but because they're easy to miss, hard to connect, and often attributed to stress or lifestyle. We built Clover because we believe that with the right information, women can walk into their doctor's office feeling prepared, confident, and heard.

What it does

Clover is an AI advocate. You tell it your symptoms, how long you've had them, and how doctors have responded. It builds you a clinical case file — a structured, evidence-based brief you can bring to your next appointment that forces the medical system to take you seriously.

The case file includes:

A clinical summary of your symptom pattern mapped to PCOS diagnostic criteria
Documented dismissals — specific language for what happened to you
A list of tests to demand by name
A paragraph you can read aloud to your doctor
Next steps you can take this week

How we built it

Clover is built with a React frontend and a Python/FastAPI backend. The AI layer uses the Anthropic Claude API to analyze symptom patterns, cross-reference the Rotterdam diagnostic criteria, and generate personalized case files. We integrated Google Places to surface nearby PCOS-friendly physicians and Google Calendar to book follow-up appointments directly from the app.

Challenges we ran into

Getting the AI to strike the right tone was harder than we expected — clinical enough to be taken seriously by a doctor, but warm enough that a woman who's already been dismissed doesn't feel like she's filling out a form. We also had to be careful not to make diagnostic claims while still being useful and specific.

Accomplishments that we're proud of

Shipping a full agentic pipeline in one night — from symptom input to a structured medical case file with real diagnostic criteria. And honestly, just finishing. It's 4am and we built something we'd actually give to people we love.

What we learned

How to build a real agentic pipeline — not just a chatbot with a system prompt, but a multi-step agent that reasons across symptoms, medical criteria, and personal history to produce something actionable. And how much it matters to build with a specific person in mind.