Eat & Train (side project)
AI-assisted nutrition and fitness coach
A personal nutrition and training assistant that helps users record meals, plan workouts, monitor progress and receive personalised recommendations without counting every calorie manually.
Try it free. Use promotional code friends — Enter it in your account settings to activate trial access.
Also available in Telegram: @eat_train_bot
Who it’s for
Designed for people managing their weight, those seeking a more mindful approach to eating, fitness enthusiasts who need structured training plans and women who want to align nutrition and exercise with their menstrual cycle.
Key features
Photo and text meal analysis
Users can photograph or describe a meal. AI estimates its composition, portion size, calories and macronutrients, which can be reviewed before saving.
AI-assisted recipe generation
Recipes are generated from available ingredients and calorie targets, with cycle-aware options and an automatically prepared shopping list.
Training plans and workout diary
Personalised training plans reflect the user’s goals, preferred frequency and available equipment. The product also includes photo-based technique analysis and a workout log.
Cycle-aware recommendations
Nutrition, recipes and training recommendations adapt to the menstrual, follicular, ovulatory and luteal phases.
AI dietitian and trainer chat
The assistant uses profile information, goals, allergies and weight trends to provide contextual answers and recommendations.
Progress and motivation
Progress charts, adjustable weight and calorie targets, daily mood and energy tracking, and activity streaks support continued use.
What makes it special
AI embedded in the workflow
Image and text analysis, recipe generation and training-plan design are integrated into the core user journeys.
Cycle-aware by design
Nutrition and training recommendations adapt to the user’s menstrual-cycle phase.
Connected journeys
Recipes feed into shopping lists and the food diary; goals influence daily recommendations; and conversations can be saved as notes.
Mobile-first
The interface is designed for quick, camera-led meal recording and convenient use during training.
My role
Designed and built Eat & Train independently, from product architecture and user journeys to monetisation. Developed the application with React and TanStack Start, added server-side rendering and integrated an AI gateway for image analysis, text recognition, recipe generation and assistant conversations. Designed the Supabase data model, row-level security and role-based access, as well as tiered subscriptions, trial access and AI usage limits. Also implemented the Telegram Web App integration, component system, onboarding, dark theme, SEO and landing-page optimisation.