NourishMe

Inspiration

What it does

NourishMe was born from the reality that one in eight Americans relies on SNAP benefits to put food on the table. For these families, the primary challenge often isn't just securing benefits, but transforming a strictly limited budget into meals that are nutritious, practical, and safe. Inspired by the Tommie Buildfest 2026 theme of AI for the Common Good and the work of partners like Loaves and Fishes and The Food Group, I decided to focus on this gap. My goal was to move beyond generic meal ideas and create a tool that respects a user’s specific budget, pantry inventory, and medical necessities like allergens. I wanted to build an AI that acts like a supportive neighbor who understands your constraints and helps you plan accordingly.

How we built it

To bring this vision to life as a solo developer, I built NourishMe as a full-stack application using Next.js 15, Supabase, and Anthropic’s Claude. The technical flow begins with a detailed onboarding process where I collect essential user context, including household size, ZIP code, dietary flags, and cooking time. This data directly shapes the meal plan generation. I implemented Claude using structured tools to ensure the AI operates within strict guardrails. To validate the results, I developed a deterministic scoring engine that evaluates every plan based on a weighted formula: 45 percent for nutrition, 35 percent for cost-effectiveness, and 20 percent for pantry utilization.

Challenges we ran into

Grounding the application in real-world data was a significant technical hurdle. I integrated a multi-source price pipeline that pulls from Kroger and utilizes Open Food Facts for barcode scanning to retrieve Nutri-Scores, Eco-Scores, and allergen data. This allows the app to generate smart grocery lists that subtract items already in the user's pantry and highlight SNAP-eligible totals. I also included a SNAP retailer locator based on ZIP codes and a day-regeneration feature that lets users swap out a single day's meal while maintaining the overall weekly budget and nutritional targets.

Accomplishments that we're proud of

The development process forced me to navigate the complex balance between hard and soft constraints. Allergens, for instance, required a zero-tolerance approach, whereas ecological preferences needed to remain flexible to avoid compromising the user's budget. I also had to ensure the interface remained simple and accessible for users who might be under significant stress. This led to the inclusion of a guest mode and clear visual indicators for budget and nutrition.

What we learned

Throughout the build, I prioritized responsible AI practices by incorporating safety disclaimers and strict rules against providing medical advice, ensuring the tool remains a supportive resource rather than a replacement for professional guidance. Through this project, I learned that creating AI for the common good requires more than just a clever model; it requires data that reflects the actual economic reality of the users. My collaboration with local nonprofits kept my focus on these real-world needs.

What's next for NourishMe

Refine the app's ability to cite specific sources and offer context-aware help, further supporting the decision-making process for families stretching every dollar.

Built With

• anthropic
• next.js
• supabase