NutriWell
Enabling personalised, socially connected nutrition through AI and digital responsibility; helping elderly citizens eat well, live well, and share healthy habits together
NutriWell reimagines how technology can support healthy ageing – turning personalised nutrition into a social and empowering experience. The project has developed five interoperable AI-driven enablers: a Nutrition Data Space, a Personal Data Wallet, an AI Nutrition Plan Generator, an AI Cuisine Allocator, and a Social Cooking Organiser. Together, these tools help users design and enjoy healthier, science-based meal plans adapted to their personal wellbeing, dietary needs, and cultural preferences – while keeping full control of their data. Combining precision nutrition science with social interaction, the platform connects AI-generated recommendations with everyday meal choices, making personalised nutrition both accurate and approachable.
The NutriWell initiative directly addresses one of Europe’s most urgent social challenges: supporting an ageing population to remain active, connected, and healthy. In Bulgaria alone, over 1.5 million people are aged 65 or older, many living on limited budgets and facing isolation or chronic dietary challenges. NutriWell responds by combining AI-powered personalisation, trustworthy data sharing, and Living Lab co-creation to build a platform that is not only smart, but human-centred.
Tested with citizen groups at SofiaLab – part of the European Network of Living Labs – NutriWell has demonstrated how AI and data spaces can be applied responsibly in real-world settings. By integrating open enablers from FIWARE, iSHARE, and dataU, the platform had ensured GDPR-compliant data use and transparent user control. The result is a modular ecosystem where citizens, dietitians, and developers alike can co-create new responsible nutrition services – from adaptive meal planning to social cooking sessions that bring people together around food and wellbeing.
WHY THIS MATTERS
Europe’s population is ageing rapidly – by 2040, nearly one in four citizens will be over 65. In Bulgaria alone, more than 1.5 million elderly people already face the combined challenges of chronic disease, limited mobility, and restricted access to affordable, nutritious food. For many, meals are repetitive, poorly balanced, or constrained by medical diets that make healthy eating feel restrictive rather than empowering. At the same time, social isolation among older adults is rising, with direct consequences for mental health and overall wellbeing.
Despite major advances in food and health technologies, a critical gap persists between personalised nutrition science and everyday life. Most existing diet and meal-planning applications are designed for highly digital or health-expert users. They tend to be overly medicalised, technically complex, or focused on data tracking rather than empowerment – a poor fit for many older adults.¹ ² ³ ⁴ Such tools often neglect cultural food preferences, overlook privacy and consent, and provide little opportunity for social connection.
These gaps reveal a deeper challenge: ensuring that digital health innovation serves all citizens, not just the most connected ones. The ageing population is diverse in its abilities, resources, and traditions – yet most digital nutrition tools assume a one-size-fits-all model. By confronting these accessibility, inclusion, and trust issues head-on, initiatives like NutriWell are vital to shaping a digital food ecosystem that is fair, inclusive, and truly human-centred.
- van der Haar, S., et al. (2023). Incorporating consumers’ needs in nutrition apps to promote healthier diets: Mixed-methods study. JMIR mHealth and uHealth, 11(1). https://mhealth.jmir.org/2023/1/e39515/
- LoBuono, D. L., & Milovich, M. (2023). Food Delivery Apps and Their Potential to Address Food Insecurity in Older Adults: A Review Nutrients, 15(17), 4402. https://www.mdpi.com/2072-6643/15/17/4402
- Kebede, A. S., et al. (2022). Digital engagement of older adults: Scoping review. Journal of Medical Internet Research, 24(12), e40192. https://www.jmir.org/2022/12/e40192/
- United Nations Economic Commission for Europe (UNECE). (2021). Ageing in the Digital Era – Policy Brief No. 26. Geneva: UNECE. https://www.un-ilibrary.org/content/papers/10.18356/27083047-26
WHAT THE SOLUTION DOES
NutriWell turns this challenge into an opportunity – showing how AI and responsible data design can make healthy eating not only simpler, but more engaging, social, and sustainable. Rather than delivering static meal plans, it connects nutrition science with human motivation: helping individuals and professionals understand how food choices align with nutritional science and wellbeing goals, offering the encouragement, flexibility, and shared experiences that make healthy habits stick.
Imagine Elena, a 72-year-old retired teacher living in Sofia. She cooks every day but lately has struggled to keep track of what her doctor recommends for her heart and blood pressure. Most nutrition apps feel overwhelming – long forms, medical jargon, passwords she can’t remember. When she hears about NutriWell, she joins a community session at SofiaLab, where the platform was co-created with seniors like her.
With a few guided steps, Elena creates her personal data profile. She enters only what she chooses to share – age, activity level, personal food preferences, and the ingredients her doctor has advised her to avoid. Behind the scenes, this information is stored securely in her Personal Data Wallet, giving her full control over who can access it. No central database, no hidden sharing – just her own trusted space for nutrition data.
Moments later, the AI Nutrition Plan Generator turns her details into a week’s menu that feels both achievable and rewarding. Behind the scenes, NutriWell’s AI modules perform detailed nutritional calculations – balancing macronutrient and micronutrient needs with personal health goals, based on validated scientific data. It balances her health goals with the foods she enjoys, automatically adjusting portions and nutrients using verified data from the Nutrition Data Space. Each meal is accompanied by simple guidance and nutritional insights – practical reminders that healthy eating can be effortless, enjoyable, personal, and socially connected rather than restrictive.
Because food is personal, Elena opens the AI Cuisine Allocator to view her plan through the lens of Bulgarian flavours. The same nutrition values now appear as local dishes: bean soup with olive oil, yoghurt with honey, and seasonal vegetables instead of imported ones. Healthy choices feel familiar, culturally grounded, and easy to maintain.
By the weekend, Elena opens the Social Cooking Organiser, where she joins a small group of others following similar plans. They schedule a cooking afternoon – part virtual, part in-person – exchanging recipes, cooking together, and comparing results. Some join from their homes, others from a local community kitchen. What began as a digital tool becomes a circle of encouragement, learning, and connection.
For the developers behind NutriWell, every one of these steps embodies a layer of digital responsibility:
- Privacy and autonomy through thePersonal Data Wallet
- Data fairness and interoperability via theNutrition Data Space
- Transparency and adaptability in theAI Nutrition Planner
- Cultural inclusion through theCuisine Allocator and
- Social wellbeing and motivation through theSocial Cooking Organiser
Together, these layers form a living example of trustworthy innovation – turning artificial intelligence into everyday motivation that helps people eat well, connect, and thrive.
HOW YOU CAN USE IT
The NutriWell project outputs are more than prototypes – they are immediately useful components for anyone working at the intersection of food, health, ageing, and digital innovation. Each audience can engage with the platform in distinct yet complementary ways.:
For citizens: NutriWell is designed first and foremost for citizens – especially older adults seeking practical, trustworthy guidance for everyday meals. It transforms nutrition planning from an often medical chore into an effortless and socially connected experience. Users create personal profiles that reflect their health needs, cultural tastes, and dietary restrictions, including allowed and forbidden ingredients. The AI-driven tools then generate personalised menus and shopping lists that respect those choices while introducing variety and balance. Through the Social Cooking Organiser, participants can stay motivated by joining cooking circles – online or in community spaces – where recipes and experiences are shared. Families can also use NutriWell together, supporting elderly relatives in developing sustainable, healthier routines.
For health professionals and care teams: NutriWell supports nutritionists, care workers, and family doctors who want to provide personalised advice while maintaining privacy and consent. Its architecture enables trusted sharing of nutrition information between citizens and care providers, supporting discussion and follow-up without exposing personal data. In Living Lab trials, healthcare and community stakeholders explored how the Personal Data Wallet and Nutrition Data Space could simplify data sharing and support citizens in taking greater ownership of their wellbeing. For professional users, NutriWell offers a suite of enablers that can be integrated into existing digital health or wellbeing platforms – such as Active@Home – to promote healthier ageing and empower citizens in managing their own nutrition.
Dietitians and nutritionists can also apply NutriWell’s AI modules to generate precision meal plans based on macronutrient and micronutrient profiles, adapting them to clinical or preventive care contexts.
For clinical trials and research: NutriWell’s data architecture and AI components can also support research and clinical trials requiring personalised yet standardised dietary protocols. The AI Nutrition Plan Generator and Nutrition Data Space enable reproducible, consent-based management of dietary interventions, making the platform a practical tool for studies in diet-sensitive conditions or behavioural health. By linking nutrition data, medical metadata, and lifestyle factors in a privacy-compliant framework, NutriWell provides a robust model for integrating responsible AI into health and nutrition research.
For developers and innovators: NutriWell’s enablers – including the Nutrition Data Space, AI Nutrition Plan Generator, and AI Cuisine Allocator – are modular components designed for reuse within the DRG4FOOD Toolbox. Developers can integrate or extend these modules in new food- and health-related platforms, adapting them for different cuisines or dietary frameworks.
The system’s open, FIWARE-based architecture ensures interoperability, while project documentation from the Living Lab demonstrates how privacy-by-design and inclusive design principles can be implemented in practice. NutriWell shows that AI-based personalisation, cultural sensitivity, and user empowerment can coexist within a single responsible digital framework.
A technical preview of NutriWell’s Identity Management and Authorization service is already accessible through a Swagger interface, illustrating how secure access and user consent are handled across the platform: Link
NutriWell has also published an initial nutrition ontology, defining how nutritional, medical, and lifestyle data can interconnect to support knowledge-based AI reasoning: Link
The project partners have committed to make the key technology enablers openly accessible through the DRG4FOOD Toolbox under appropriate open-licence terms. Links to the corresponding repositories and documentation will be provided here as they become available.
DIGITAL RESPONSIBILITY IN PRACTICE
The NutriWell team used the Digital Responsibility Goals (DRGs) as both a design compass and an evaluation framework throughout the project. Each of their AI-based enablers was developed using the principles of trustworthy-by-design AI, privacy-by-design data management, and inclusive Living Lab co-creation. By aligning its architecture with European standards such as GDPR, ALTAI, FIWARE, and iSHARE, and by connecting with broader ecosystem initiatives including i4Trust, FOODITY, and GAIA-X, the project demonstrates how responsible AI and data-space technologies can be applied concretely within the food, health, and ageing domains.
From the project’s outset, elderly citizens, care workers, and healthcare professionals took part in Living Lab workshops at SofiaLab, where early prototypes of the NutriWell platform were introduced and tested in realistic settings. Participants shared feedback on clarity, usability, and motivation. Sessions combined technical demonstration with comprehension: users learned how their information is stored, how consent works in the FIWARE-based architecture, and how AI-generated nutrition plans can be adjusted to their own needs.
Short orientation modules and guided exercises helped participants to recognise everyday digital actions such as granting or withdrawing data access, checking plan recommendations, or joining group-cooking activities online. These activities were designed to build confidence and understanding among people who often describe themselves as “non-digital,” turning digital literacy into a practical, empowering experience rather than an abstract skill. By opening the design process to end users and other key stakeholders, NutriWell transformed technology testing into shared learning – proactively demystifying data, personalisation, and consent.
Useful resources for Living Lab and Co-Creation approaches can be referenced at the European Network of Living Labs (ENoLL) Knowledge Hub of which SofiaLab is a member.
From the Digital Responsibility Database, the Design Councils’ Framework for innovation, Double diamond model for iterative, user-centred design and Design methods for developing services as well as IDEO’s Design Kit provide additional related resources.
NutriWell applies a federated, standards-based security model across their enablers with identity and authorisation managed through FIWARE KeyRock, which issues OAuth 2.0 tokens and JSON Web Tokens (JWT) to verify users and control access between distributed components such as the Nutrition Data Space and Personal Data Wallet. Each module authenticates requests locally but under a unified token scheme, helping to create a seamless yet compartmentalised security layer. All data traffic between components uses HTTPS and IPsec encryption, while internal routing enforces role-based permissions so that only authorised services can access or exchange information. This integration with the iSHARE and dataU frameworks aligns NutriWell with the security and trust rules emerging for European data spaces.
What makes NutriWell distinctive is how these established security standards are applied in a citizen-friendly way. Users log in through a single secure interface, while the underlying system manages encryption, token exchange, and authorisation. This combination of distributed identity management, token-based access, and consent-aware data sharing demonstrates how robust cybersecurity frameworks can operate transparently and accessibly for non-technical audiences.
Privacy-by-design lies at the core of NutriWell’s architecture, treating data protection as a system feature rather than a compliance afterthought. Personal information is stored in the Personal Data Wallet Adapter, a distributed service connected to DataU – a decentralised, GDPR-compliant data-sharing framework. DataU’s architecture (dashboardU, nodeU, proxyU) enables explicit, dynamic, and revocable consent, ensuring that users remain in control of their information at all times. No personal data is ever stored centrally: information remains with the user and is exchanged peer-to-peer through encrypted channels once consent is verified.
Health profiles and dietary preferences stay local to the wallet, while sensitive information is pseudonymised or anonymised before analytic use, and only aggregated data enter the Nutrition Data Space. Access and reuse are governed by explicit, revocable consent; and integration with the iSHARE Trust Framework ensures that identity, authorisation, and consent policies are applied consistently across all NutriWell enablers and any connected services.
Together, DataU and iSHARE provide the governance and technical mechanisms that allow consent and access permissions to travel with the data, making each transaction privacy-aware by default.
The architecture essentially enforces privacy-by-design: every analytic dataset is detached from personal identifiers, and every service is insulated from unnecessary exposure. Development followed the principles of the General Data Protection Regulation (GDPR) as well as the Assessment List for Trustworthy AI (ALTAI), which guided risk assessment, data minimisation, and lawful processing. By embedding these safeguards into its technical foundation, NutriWell shows how AI-enabled food health services can remain compliant and respectful of personal autonomy without adding friction for users.
Further information on DataU and its integration with European data-space frameworks such as iSHARE can be found via the FOODITY project documentation and Jibe Company and at the iSHARE Foundation
NutriWell promotes data fairness by ensuring that health and nutrition information can be shared and reused safely, transparently, and on equal terms.
Its Nutrition Data Space follows thae FAIR principles – Findable, Accessible, Interoperable, and Reusable – and is built on FIWARE context-broker technology to avoid vendor lock-in and enable cross-domain integration.
To guarantee semantic consistency, the project combines open vocabularies such as FoodOn with its own published nutrition ontology, aligning dietary and medical concepts in a machine-readable way. This approach allows anonymised, consent-based data from the Personal Data Wallet and other sources to be reused for research or public innovation while preserving user ownership.
By applying shared standards, open APIs, and interoperable semantics, NutriWell turns fairness into practice – ensuring that data benefits citizens, developers, and researchers alike without sacrificing trust or control.
From the DRG Database; the FAIR data principles and practical guidance through the GO FAIR Initiative and the FIWARE Foundation.
Trustworthy AI principles guide the development of NutriWell’s intelligent components, ensuring that algorithmic decision-making remains transparent, fair, and accountable. The AI Nutrition Plan Generator and AI Cuisine Allocator were developed in alignment with the Assessment List for Trustworthy AI (ALTAI), which guided checks for data quality, bias, and proportionality of recommendations. All nutritional data originate from verified scientific sources – EFSA, FooDB, and Edamam – ensuring that algorithmic outputs rest on transparent and authoritative datasets.
Explainability is built into the user experience: meal recommendations can be inspected and adjusted, showing the reasoning behind substitutions or portion changes. Evaluation against diverse age, gender, and dietary profiles helped detect bias and improve robustness, while internal documentation captures data lineage and model-training parameters for auditability.
By applying ALTAI within a practical health-nutrition setting, NutriWell demonstrates how small projects can translate European AI ethics guidelines into tangible development routines. More information on the ALTAI framework can be found here.
Transparency in NutriWell extends from system architecture to user experience. Every AI-generated recommendation can be inspected and adjusted, allowing citizens to understand how and why meal plans are suggested. The Nutrition Data Space maintains traceable data flows between enablers, with each transaction logged and consent-verified through iSHARE. Open documentation – including NutriWell’s published nutrition ontology – ensures that data structures and reasoning models are visible to developers and researchers.
Within the SofiaLab Living Lab, participants discussed how their information was collected and reused, turning transparency into two-way communication rather than a static disclosure. By combining explainable AI with traceable data flows, NutriWell demonstrates that transparency can be both technical and human.
NutriWell reinforces human agency by ensuring that technology supports rather than replaces personal decision-making. Within the platform, citizens retain ownership of their data, choosing what to share, with whom, and for how long. AI-generated meal plans and recommendations remain fully advisory – users can view, adjust, or reject them at any time, maintaining control over dietary and lifestyle choices.
Participatory development in the SofiaLab Living Lab placed elderly users and healthcare professionals directly in the design loop, so that system behaviour reflected human priorities such as comfort, clarity, and self-determination.
By coupling data sovereignty with explainable, user-adjustable AI, NutriWell shows how digital food health tools can strengthen rather than diminish individual agency and identity.
CONTRIBUTION TO THE TOOLBOX
The NutriWell project contributes a modular framework of AI-driven and data-centric components designed to demonstrate how responsible digital design can be applied in the nutrition and wellbeing domain. Each component translates one or more Digital Responsibility Goals (DRGs) into practice, combining privacy-by-design data handling, trustworthy algorithm design, and socially connected user interaction.
Together, these elements form an interoperable architecture that is being prepared for inclusion within the DRG4FOOD Toolbox under open-access conditions. Each will be accompanied by implementation guides, architecture diagrams, and API documentation to support understanding and future reuse.
- Nutrition Data Space (Generic Enabler): FIWARE-based data infrastructure that models FAIR, consent-aware exchange of anonymised nutrition and wellbeing data, aligning with European data-space principles. Expected Toolbox assets: reference architecture diagram, configuration templates, and FIWARE integration documentation. (Link to repository pending)
- Personal Data Wallet Adapter (Specific Enabler): Decentralised data-storage and consent-management interface based on DataU, demonstrating how individuals can dynamically manage access to their personal information. Expected Toolbox assets: adapter code repository, consent-flow documentation, and API reference for DataU integration. (Link to repository pending)
- AI Nutrition Plan Generator (Specific Enabler): Algorithmic engine that produces personalised nutrition plans from validated scientific datasets (EFSA, FooDB, Edamam) with user adjustment and oversight features. Expected Toolbox assets: trained model and example dataset (potentially on Hugging Face), model-training scripts, and evaluation notebook. (Link to model repository pending)
- AI Cuisine Allocator (Specific Enabler): Cultural-adaptation module translating nutritional recommendations into familiar, region-specific meals and ingredients. Expected Toolbox assets: algorithmic logic or rule base, sample datasets for regional cuisines, and integration examples for connecting to the AI Plan Generator. (Link to repository pending)
- Social Cooking Organiser (Specific Enabler): A collaborative module supporting group-cooking sessions and social engagement, linking individual wellbeing to community participation. Expected Toolbox assets: demonstrator web or mobile application, backend source code repository, and user-onboarding guide. (Link to demonstrator pending)
All enablers communicate through a FIWARE context-broker architecture with iSHARE-compliant authorisation, aligned with i4Trust and eIDAS frameworks to ensure secure, traceable and interoperable data exchange across systems. An OpenAPI/Swagger interface for NutriWell’s FIWARE KeyRock identity- and authorisation-management layer is publicly available, offering insight into the project’s secure-access architecture, while a publicly available nutrition ontology (r2.owx) further illustrates this design.
By providing a combination of architectural, AI, and participatory components, NutriWell contributes both practical and conceptual assets that connect personal data sovereignty, responsible AI, and food-system innovation within one interoperable framework.
IMPACT AND OUTLOOK
The NutriWell project has shown how responsible digital innovation can address one of Europe’s most pressing health and social challenges: supporting healthy ageing and nutritional wellbeing in later life. Through the SofiaLab Living Lab, the team co-created and tested digital nutrition tools with more than 120 elderly participants, healthcare professionals, and community stakeholders across Bulgaria. Early trials demonstrated that seniors could meaningfully engage with personalised meal-planning tools when interfaces were simple, consent flows transparent, and recommendations culturally familiar.
Participants reported improved understanding of nutritional balance, greater confidence in using digital services, and stronger motivation to maintain healthier habits. Healthcare partners highlighted potential efficiency gains: the Personal Data Wallet and Nutrition Data Space prototypes reduced administrative effort in sharing nutrition plans and wellness data, while maintaining data sovereignty and compliance with the GDPR.
Beyond individual outcomes, NutriWell contributes to broader system learning, demonstrating how AI, data spaces, and consent-based architectures can interoperate to deliver citizen-controlled nutrition services that remain technically robust and socially inclusive. Its modular components – spanning data infrastructure, personal consent management, and AI-driven meal generation – promise to offer reference models for future responsible-by-design food-tech solutions.
Looking ahead, the NutriWell partners plan to deepen collaboration with health and nutrition authorities in Bulgaria and neighbouring regions, exploring integration of the platform into preventive health and community-care programmes, while contributing to European initiatives such as Food 2030, GAIA-X, and the European Health Data Space.
Future development will focus on refining the AI Nutrition Plan Generator, expanding datasets for additional dietary frameworks, and aligning the Data Wallet with emerging European interoperability standards.
Through these next steps, NutriWell provides a concrete path from Living Lab experimentation to operational service design.
QUICK FACTS
- Funding
- DRG4FOOD Open Call #1
- Use case
- Targeted Nutrition
- Partners
- Start date
- May 2024
- End date
- Dec 2024
- Resources