MatthewEatBot — AI nutrition assistant in Telegram

Development of a multifunctional Web App for automating food diary logging, nutrient analysis, and real-time health progress visualization.

Try the bot on Telegram

The main objective was to eliminate the "routine entry" barrier, which often causes people to stop keeping a food diary. We had to create a tool that simplifies calorie tracking as much as possible, visualizes user progress, and provides "smart" dietary adjustment tips based on data.

Project Sector — HealthTech, FoodTech, lifestyle apps.
Technologies — React.
Implementation Period — 3 months.

The main feature of MatthewEatBot is the seamless integration of the Whisper voice model, which allows the user to simply dictate their meal. The system automatically recognizes products, matches them with the database, and calculates calories and macronutrients (PFC), eliminating the need for manual catalog searching. Furthermore, the built-in AI-based recommendation system analyzes user habits and provides personalized tips for dietary improvement.

We implemented a hybrid architecture, combining the capabilities of a bot and a Mini App. A user-friendly interface with progress charts and nutrition history was deployed within Telegram. To ensure data accuracy, we integrated external APIs and configured a Python-based queue system for processing voice messages. This allowed us to create a product that works "on the fly," even with an unstable internet connection.

MatthewEatBot has become a comprehensive tool for mindful health management, completely eliminating the complexity of data entry. The product demonstrated high user engagement thanks to the simplicity of "on-the-go" logging and clear visual statistics. The successful launch established a solid technical foundation for scaling the service across the CIS market and adding new HealthTech features.