This article outlines the development journey of SOSfood Consumer APP, a consumer-oriented application, progressing from the conceptual stage (TRL 2) to demonstration in an operational environment (TRL 6). It details the technical phases, validation methods, and key outcomes, emphasizing the app’s potential impact on user experience and readiness for market scaling.

SOSFood project is dedicated to advancing sustainability in food systems by leveraging data-driven insights, innovative technologies, and multi-stakeholder collaboration. Within this vision, the Consumer APP plays a central role empowering user through interactive and intelligent tools that visualize their food consumption impact and guide them with personalized recommendations such as better choices or useful recipes.

The project has a duration of four years, but in this timeframe, the first deliverable of the consumer APP it was expected within the end of the first year. So, from the very beginning, an iterative approach was adopted, with the goal to ensure a smooth transition from an initial concept to a validated prototype ready for real-world testing.

Methodology

The development process was structured into five stages aligned with TRL progression:

  • TRL 2: Concept formulation and functional specification.
  • TRL 3: Experimental proof of concept through a minimal viable product (MVP).
  • TRL 4: Laboratory validation, including usability and performance testing.
  • TRL 5: Integration with external systems and beta testing in a relevant environment.
  • TRL 6: Demonstration in an operational setting with real users.

Stage 1: The Spark (TRL 2)

It all started with a question: How can we empower people to make food choices that are good for them and for the environment?

Through co-creation workshops and user research, we shaped a vision: an app that combines science, AI, and design to deliver personalized guidance. We mapped out the core features—eco-health fingerprint visualization, product scanning, smart recommendations, and a recipe hub—and defined the technologies to bring them to life.

At this stage, the app’s vision and functional architecture were defined, leading to some key outcomes:

  • Modular Architecture: Eight core modules, including Eco-Healthy Footprint Visualization, Product Scanning, AI-based Recommendation Engine, Personalized Recipes, and Offline Features.
  • Functional Requirements: Personalized eco-health fingerprint, product scanning, AI-driven recommendations, multilingual support, and WCAG 2.1 accessibility compliance.
  • Technology Choices:
    • Frontend: Flutter (Dart) or React Native for cross-platform mobile development.
    • Backend: Node.js with Express for scalable, asynchronous operations.
    • Database: MySQL for structured storage of user profiles, recipes, and ecological metrics.
    • AI Integration: TensorFlow Lite or ONNX for on-device inference; rule-based fallback logic.
    • External APIs: OpenFoodFacts, EcoScore, EU nutrition datasets.

Stage 2: First Steps (TRL 3)

Next came the Minimum Viable Product (MVP). A simple prototype, but powerful enough to prove the concept:

  • A radar chart showing your health and sustainability scores.
  • A barcode scanner to capture product data.
  • AI-driven suggestions for better alternatives.

Even in this early stage, the app felt different—interactive, intuitive, and designed for real impact.

This first prototype was developed to validate core functionalities:

  • Eco-Healthy Fingerprint Visualization: Interactive radar chart showing health, environment, and sustainability scores.
  • Product Scanning: Barcode recognition via libraries like ZXing or ZBar, integrated with external databases.
  • Basic Recommendation Engine: Hybrid AI model providing alternative products and menus.
  • Offline Mode: IndexedDB or SQLite for local caching and deferred synchronization.
  • UI/UX: Minimalist design, responsive layout, and onboarding tutorial.

Stage 3: Refining the Experience (TRL 4)

Usability tests, performance checks, and accessibility reviews shaped every detail. The feedback was clear: users loved the clean design and personalized insights. We fine-tuned the interface, optimized speed, and ensured compliance with WCAG 2.1 standards. Security was a priority too—encrypted data, JWT authentication, and GDPR compliance.

More in details, controlled testing focused on usability and performance:

  • Usability Tests: Mockups validated by consortium members; average UI rating of 4.1/5.
  • Performance Metrics:
    • Frontend load time < 3 seconds.
    • API response time < 300 ms.
  • Accessibility Compliance: WCAG 2.1 AA standards verified.
  • Security: JWT-based authentication, bcrypt password hashing, and input validation with Joi.

Stage 4: Opening the Doors (TRL 5)

Integration was the next big leap. The app was conected to OpenFoodFacts, sustainability datasets, and our curated recipe repository. Beta testers explored the app in real-world conditions, scanning products, comparing footprints, and discovering menus tailored to their culture and preferences. Their feedback drove improvements in onboarding, clarity, and engagement.

From a technical point of view, the app was integrated with external systems and APIs and tested in relevant environments:

  • Data Synchronization: Offline-to-online sync using UUIDs and timestamps.
  • Recipe Repository: Expert-validated recipes with nutritional and ecological metadata.
  • Beta Testing: Conducted with real users to refine onboarding, improve scorecards, and clarify eco-impact metrics.
  • Deployment Strategy: Docker-based containerization for frontend, backend, and database; orchestrated via Docker Compose.

Stage 5: Ready for the World (TRL 6)

The app is now prepared for real-world demonstration:

  • Full Feature Set: AI-powered recommendations, personalized menus, multilingual support, and motivational engagement tools.
  • Operational KPIs:
    • System uptime ≥ 99.5%.
    • Offline sync success rate ≥ 98%.
    • Automated backups every 24 hours.
  • Continuous Improvement: Feedback loops and KPI dashboards (Grafana) for monitoring performance and user satisfaction.

SOSFOOD wants to be more than an app: a personal sustainability coach in your pocket.

  • AI-powered recommendations help swap products for healthier, greener options.
  • Personalized menus and recipes make sustainable eating delicious and practical.
  • Offline functionality ensures access anywhere, even in rural areas.

Technologies employed included Flutter for cross-platform performance, Node.js for scalability, MySQL for data integrity, and TensorFlow Lite for smart AI, supported by iterative UX research and agile development practices.

Impact – A catalyst for change

The SOSFOOD App is more than a digital tool, it’s a catalyst for change. By turning complex sustainability metrics into clear, actionable insights, it empowers individuals to make informed choices every day. Each scan, each recipe, each recommendation contributes to reducing carbon footprints, promoting healthier diets, and supporting local food systems. The impact extends beyond users: it fosters a culture of responsibility and transparency across the entire food value chain.

What’s next?

The journey doesn’t end here. The next phase focuses on scaling adoption, integrating real-time data from food producers, and enhancing the AI engine for even smarter recommendations. We aim to expand multilingual support, strengthen interoperability with other sustainability platforms, and introduce gamification features to boost engagement.

The vision is clear: to make sustainable eating the norm, not the exception. With SOSFOOD, we’re building a future where technology and responsibility go hand in hand—one meal at a time. Future work will focus on achieving TRL 7–9.

The article was prepared by Raffaele Bini | DIH Datalife.