AllerNavi

Inspiration and What it does

I built AllerNav for students and anyone managing food allergies who needs to make quick dining decisions without reliable information. Most restaurant apps optimize for ratings, price, or distance, but not for allergy risk. That forces people to manually search menus and dig through reviews every time they want to eat, which is stressful and time-consuming.

AllerNav takes an allergy-first approach. It helps users compare nearby restaurants by analyzing review text, surfacing relevant safety signals, highlighting possible risks, and pointing users toward menu items that may be worth verifying. The goal is not to replace judgment, but to make the research process faster, clearer, and less overwhelming.

How I used Claude / AI

I used AI in two ways: as a development partner and as part of the product experience itself.

During development, Claude helped me refine the product direction, think through edge cases in allergy-related language, and improve how I framed trust and safety in the user experience.

Inside the app, AI helps turn messy restaurant data into useful decision support. AllerNav scans review text for signals like staff knowledge, cross-contact concerns, allergic reaction reports, and signs of accommodation, then generates a short decision brief and suggests menu items to verify. I also built in a heuristic fallback path so the experience still works even when the model is unavailable.

Challenges we ran into

The biggest challenge was avoiding false confidence. An AI system could make a place sound safer than it really is, miss cross-contact issues, or recommend an item that appears safe but is not. In a problem space like food allergies, that kind of mistake matters.

To address that, I designed AllerNav to be evidence-backed rather than guarantee-based. Instead of labeling a restaurant as simply “safe,” the app shows confidence levels, caution flags, and real review excerpts so users can understand why a place appears promising or risky.

I also constrained the recommendation layer so it only suggests items that already exist on the provided menu, and I use structured outputs plus non-AI fallbacks when the model is unavailable. The product is meant to support safer decisions, not replace direct confirmation with the restaurant.

Accomplishments that we're proud of

We’re proud that AllerNavi focuses on a real, stressful problem that most restaurant tools ignore. Instead of ranking places by popularity alone, we built an allergy-first experience that helps users quickly understand which restaurants are worth a closer look.

We’re also proud of how we handled trust. Rather than presenting AI as a perfect answer, the app surfaces confidence levels, caution flags, and review evidence so users can understand why a place looks promising or risky. On the technical side, we combined restaurant search, allergy-aware scoring, menu-based recommendations, and fallback heuristics into one simple product that works without requiring an account.

What we learned

One of the biggest lessons from building AllerNav was that trust matters as much as functionality. In a safety-related product, it is not enough for the system to be helpful. Users also need to understand the limits of what the system knows and why it is making a recommendation.

We also learned that AI is most useful here when it acts as a decision-support layer, not as an authority. Combining AI with visible evidence, structured constraints, and fallback logic led to a much more responsible and practical experience than relying on a model alone.

What's next for AllerNavi

Next, we want to make AllerNav more personalized, reliable, and community-driven. That includes adding verified user-submitted allergy experiences, improving how we weigh newer reviews, and tailoring results more closely to different allergens and risk tolerances.

We’d also like to build a stronger decision-support workflow around the search itself, including a “call ahead” checklist, more detailed cross-contact guidance, and richer menu verification. Longer term, we see AllerNav becoming a practical safety companion for anyone navigating food allergies in unfamiliar places.

Built With

• api
• codex
• google
• node.js