NOVA - AI-Powered Nutrition Intelligence for Personalized Health and Wellbeing

N.O.V.A. - Nutrition Optimization and Vitamin Advisor

NOVA is available as a live technology demonstrator developed by INNOLABS – ICI Bucharest.

👉 Explore the platform at https://nova.innolabs.ro/

The demonstrator showcases how artificial intelligence, nutritional science, semantic search, and agentic systems can be combined to create an intelligent assistant capable of understanding foods, recipes, dietary habits, and nutrition goals in natural language.

---

Making Nutrition Data Accessible Through Artificial Intelligence

Nutrition information is everywhere, yet understanding how food choices affect our health remains difficult for most people. Nutritional labels, food composition databases, calorie calculators, and dietary recommendations often require significant effort to interpret and apply in everyday life.

N.O.V.A. was created to bridge this gap.

Developed within INNOLABS, N.O.V.A. is an AI-powered nutrition intelligence platform that transforms complex nutritional datasets into actionable knowledge through conversational interfaces and intelligent reasoning. Rather than requiring users to search through tables or manually calculate nutritional values, N.O.V.A. enables natural interactions such as:

• “How much protein is in 200 grams of chicken breast?”
• “Compare salmon and tuna from a nutritional perspective.”
• “Create a high-protein meal plan for the next week.”
• “Suggest healthy alternatives to white rice.”
• “Analyze my recipe and calculate its nutritional values.”

By combining verified food composition databases with advanced AI technologies, N.O.V.A. provides users with reliable, explainable, and personalized nutritional insights.

From Ingredients to Nutritional Intelligence

One of the most powerful capabilities of N.O.V.A. is its ability to understand complete recipes and estimate their nutritional composition automatically.

Consider the following request:

“Give me the main nutritional values for these Vietnamese summer rolls…”

followed by a detailed list of ingredients such as rice paper wrappers, lettuce, cucumber, radishes, ginger, carrots, sesame seeds, olive oil, and miso paste.

Instead of requiring users to manually calculate calories, proteins, carbohydrates, fats, vitamins, and minerals, N.O.V.A. can:

• Identify individual ingredients.
• Estimate ingredient quantities and edible portions.
• Retrieve nutritional information from structured food databases.
• Aggregate nutritional values across the entire recipe.
• Calculate per-serving nutritional breakdowns.
• Highlight key nutrients and dietary characteristics.
• Explain nutritional benefits in natural language.

This capability transforms recipes into actionable nutritional knowledge, making it easier for users to understand the health impact of their meals.

AI Agents that Understand Food

At the heart of N.O.V.A. are intelligent AI agents specifically designed to reason about food and nutrition.

Unlike traditional chatbots that rely exclusively on language model knowledge, N.O.V.A. combines conversational AI with structured nutritional databases and specialized retrieval systems.

When a user asks a question, the platform can:

1. Understand the user’s intent.
2. Identify foods, ingredients, recipes, or dietary goals.
3. Retrieve relevant nutritional information.
4. Perform calculations and comparisons.
5. Generate human-readable explanations.
6. Provide personalized recommendations.

This architecture allows the system to remain grounded in verifiable nutritional data while offering the flexibility of natural conversation.

Semantic Search for Food Discovery

Food terminology varies enormously across regions, cultures, and personal preferences.

A user searching for “aubergine,” “eggplant,” “vinete,” or a local recipe may expect the same underlying ingredient to be recognized.

To address this challenge, N.O.V.A. employs semantic search technologies powered by vector embeddings and artificial intelligence. Rather than relying solely on exact keyword matching, the platform understands the meaning and context of food-related queries.

This enables users to discover foods, ingredients, and alternatives using everyday language, even when the exact terms do not exist in the database.

Personalized Meal Planning

Nutrition is highly personal.

Different individuals have different goals, dietary restrictions, allergies, activity levels, and health considerations.

N.O.V.A. incorporates intelligent meal planning capabilities that allow users to generate personalized meal plans tailored to their specific needs.

The system can consider:

• Daily caloric targets.
• Protein, carbohydrate, and fat requirements.
• Vegetarian and vegan diets.
• Food allergies and intolerances.
• Weight management goals.
• Sports nutrition requirements.
• Lifestyle and meal scheduling preferences.

By combining nutritional science with AI reasoning, N.O.V.A. can generate meal plans that are both practical and nutritionally balanced.

Explainable and Responsible AI

Trust is essential when AI systems are used in domains related to health and wellbeing.

For this reason, N.O.V.A. was designed around principles of transparency and explainability.

Recommendations generated by the platform are grounded in structured nutritional datasets rather than relying exclusively on generative AI outputs. Users can inspect nutritional values, understand how recommendations are produced, and receive evidence-based explanations for dietary suggestions.

This approach helps reduce misinformation and promotes informed decision-making.

A Research Platform for Future Digital Health Solutions

Beyond its immediate applications, N.O.V.A. serves as a research and innovation platform for exploring the future of AI-assisted nutrition and digital health.

The technologies developed within the project have potential applications in:

• Preventive healthcare.
• Personalized nutrition.
• Wellness and fitness platforms.
• Digital therapeutics.
• Healthcare virtual assistants.
• Public health initiatives.
• Nutrition education.
• Smart food recommendation systems.

The project demonstrates how large language models, semantic search, structured knowledge bases, and agentic AI workflows can be integrated into trustworthy decision-support systems for everyday use.

Advancing AI Innovation at INNOLABS

N.O.V.A. reflects INNOLABS’ commitment to developing practical applications of artificial intelligence that generate measurable societal impact.

The project combines expertise from multiple research domains, including:

• Artificial Intelligence and Agentic Systems.
• Natural Language Processing.
• Semantic Search and Knowledge Retrieval.
• Data Analytics.
• Healthcare Technologies.
• Decision Support Systems.

By transforming nutritional data into accessible knowledge and personalized guidance, N.O.V.A. illustrates how AI can empower individuals to make better-informed dietary choices while laying the foundation for future innovations in digital health and wellbeing.