Inspiration
Dining should be enjoyable, not a guessing game.
- Whether someone has severe allergies, mild food sensitivities, or dietary preferences, or is just trying to eat intentionally, it’s hard to know where it's truly safe and comfortable to eat.
- Traditional review platforms focus on taste & vibes, not food safety, trust, or allergen awareness.
We wanted to build a tool that helps you choose confidently, whether you’re: Avoiding sesame or dairy Eating vegan or halal Traveling and unsure about food risks Dining out with friends who all have different needs
Food should bring people together, not create stress.
What it does
AllerNav is an intelligent dining assistant that:
- Scores restaurants by allergy-friendliness & safety
- Scores menu items individually (dish-level analysis)
- Lets users select personal allergy profiles
- Shows safety on a map UI
- Collects crowdsourced feedback with trust weighting
- Provides manual review insights & restaurant info when APIs fail
- Gives a trust score bar for each restaurant
- Filters places by vibe (family-friendly, study spot, late-night, etc.)
Think of it as Yelp + safety + personalization.
How we built it
We built AllerNav using Streamlit for the interface and Folium for the map. Backend logic uses Python scoring functions, CSV files for persistent storage, and session state to handle user feedback. The design originally supported live API calls for restaurant metadata and allergy-aware text generation, but due to API key limits, we added graceful fallback data to ensure a smooth real-world demo experience. We implemented rule-based allergen inference, cross-contact penalties, and community feedback weighting to generate dynamic safety scores. A “sensitivity level” control adjusts risk tolerance to make the system useful for both highly-sensitive users and people just trying to eat consciously.
Challenges we ran into
The hardest part was designing a scoring system that feels trustworthy without overwhelming users. Food safety varies by cuisine, preparation method, and the reliability of staff knowledge—and not all data is structured or consistent. We also ran into issues with external APIs, map click behavior, and keeping the UI clean while displaying meaningful information. To handle this, we created robust fallbacks and simplified interactions. Balancing clarity, safety, and usability was a real challenge.
We initially planned to use Google Places API, SerpAPI, Gemini AI for NLP food safety summaries but we replaced it with local demo stubs due to time & API issues
Accomplishments that we're proud of
We built a fully working prototype that feels useful and intuitive. We have interactive maps, live scoring, review-based community voting, a safety meter, a clean UI, and restaurant detail pages. We successfully handled API failures by building our own reliable offline data layer. We're proud that AllerNav is not just for allergy sufferers—it helps anyone make more confident dining decisions.
What we learned
We learned how hard real-world food decision making is, and how valuable personalization is. We learned to design fallback systems when APIs are unreliable. We improved our UX skills by designing interfaces that simplify complex safety reasoning. Most importantly, we learned how much peace of mind matters in food choices and how technology can support that.
What's next for AllerNav
After the hackathon, we want to expand AllerNav to support:
- Real-time API-powered restaurant and menu intelligence
- Verified restaurant profiles and staff-training badges
- Travel mode (“safe places near me” anywhere in the world)
- Social trust systems and reviewer badges
- Restaurant dashboards to encourage allergen-awareness
- AI-powered menu scanning and cross-contact alerts
Our goal is to make dining safer, more inclusive, and more confident for everyone, whether you're severely allergic, mildly sensitive, or just eating intentionally.
Log in or sign up for Devpost to join the conversation.