Inspiration
Supplement use is rapidly growing, but users often rely on conflicting advice and marketing claims, making it hard to know what’s actually safe or effective.
Studying NLP and being interested in AI evaluation and source searching is what brought us this far, propelling us to create a robust and flexible algorithm that minimises the risks of any ai agent hallucinating during the process.
What it does
gramWIN helps users verify supplement claims, understand evidence, and make informed decisions through clear, personalized health insights.
How we built it
We built a multi-agent AI pipeline that breaks down claims, retrieves and evaluates evidence, analyzes contradictions, and outputs a scored verdict with explanations.
Challenges we ran into
Balancing medical accuracy with simple, digestible outputs was difficult, especially when dealing with conflicting or inconclusive research.
Accomplishments that we're proud of
We translated complex health evidence into intuitive insights while maintaining transparency, helping users move away from blind trial-and-error.
What we learned
We learned that trust in health tech comes from clear explanations and balanced reasoning, not just raw data or AI outputs.
What's next for gramWIN
We plan to expand personalization, improve long-term tracking, and integrate more high-quality medical data to strengthen reliability and user trust.
Project Link to Elevator Pitch Slides
https://canva.link/9140js4ujxy5svk
Repository Link to GitHub
Built With
- expo.io
- fastapi
- python
- react
- typescript
Log in or sign up for Devpost to join the conversation.