PrepAi — AI meal planning web application

The brief

The client needed a web application that could generate meal plans tailored to individual health goals, dietary restrictions, and macro targets — without the friction of manual calorie counting or meal research. The existing market had either rigid template plans or overly complex macro calculators that required nutrition expertise to use effectively.

The goal was to build a tool that felt intuitive for everyday users while remaining accurate enough for coaches managing client nutrition. The application needed user accounts, goal tracking, plan history, and a tiered subscription model to serve both individuals and professional coaches.

An additional requirement emerged during discovery: users wanted quick answers to common meal questions ("What can I snack on?", "How do I hit protein targets?") without waiting for full plan generation. This led to the Nutrition Nancy chat feature — an AI assistant that provides practical, plain-language guidance on meals, snacks, and portion decisions.

The build

The application was built as a custom web platform with user authentication, secure payment processing, and AI integration. The meal plan generator uses structured prompts to Claude (Anthropic's AI) with validated output formatting — ensuring plans are accurate, varied, and aligned with user goals rather than generic recipe dumps.

User accounts support role-based access: free users get basic plan generation, Pro users unlock advanced features and unlimited plans, and Coach accounts can manage multiple client profiles with tracked history. Each plan is saved to the database so users can reference past weeks or iterate on preferences without starting from scratch.

The Nutrition Nancy chat feature was integrated as a conversational interface using the same AI backend, optimised for quick, actionable answers rather than detailed meal plans. Context is maintained per session so follow-up questions stay relevant, and all chat history is logged for user reference.

Subscription management was built using Stripe, with webhook handling for plan upgrades, downgrades, and cancellations. The application tracks active subscriptions and automatically adjusts feature access without requiring manual intervention. Tech stack: PHP (backend logic), Tailwind CSS (responsive UI), Claude API (AI meal generation and chat), Stripe (payments), MySQL (user data and plan storage).

The result

PrepAi launched in production with user registration, tiered subscriptions, goal-based meal plan generation, and the Nutrition Nancy chat coach. Users can generate personalised meal plans in under 2 minutes, save plan history, and get quick answers to meal decisions through the chat interface.

The Coach tier enables professional nutritionists and trainers to manage multiple client profiles, track their meal plan history, and adjust goals as client needs evolve. The subscription model supports sustainable development with recurring revenue from Pro and Coach plans, while the free tier provides a low-friction entry point for new users.

All code and infrastructure are owned by the client, hosted on managed infrastructure with automated backups and uptime monitoring. The application continues to receive feature updates and performance improvements as usage grows.