From Sensor to Flavour — How We Built VITALIJA No. 01

Before VITALIJA No. 01 was a bottle concept, it was a question beneath a pergola: what was happening to the fruit, and which moment of the season should we keep?

The work began at Sarajevo Hills with a mature vine known in the project as Vitalija. Sunlight, shade and heat did not reach every leaf and bunch in the same way. Following those differences became a practical study of how a growing plant could inform a drink.

This is a development notebook, not the announcement of a fully validated finished product. It follows the 2026 field record and subsequent Aroma Lab bench observations through the current verjus processing baseline.

One vine, several microclimates

A pergola creates an uneven landscape within a single canopy. Outer bunches meet direct exposure; others sit deeper inside the foliage. Handheld Fluke infrared measurements made that contrast visible.

On 16 August at 16:00, the field record logged south-west bunches at 36–39°C, interior bunches around 33°C and north-east bunches around 32°C. North-east leaves were recorded at 28–30°C. These were surface readings at a particular time, not continuous measurements of the berry interior.

For the grower, the implication was a better question: when we say the grapes are ready, which part of the vine are we describing? The observations suggested heterogeneity worth following. They did not establish a causal relationship between a particular temperature and a particular flavour.

From sensing to a useful record

°Brix and pH were tracked as the season progressed. IoT data recorded through ThingSpeak added a longer environmental timeline. The supplied export contained 6,268 records between 18 August and 27 September, with its first field described in the project diary as soil temperature.

That record is valuable, but its limits matter. Sensor position, depth and calibration need documentation before absolute readings can support physiological conclusions. The handheld measurements also need a more standardized protocol in future seasons.

WineTwin is the digital-twin and R&D framework around this vineyard work. It connects the ambition to follow environmental history with fruit chemistry, sensory observations and product decisions. MATLAB belongs to that analytical direction; we are not presenting a validated predictive model or claiming that software selected the 2026 harvest automatically.

This is a map of the research workflow. It does not imply that every analytical link has been validated or that every instrument measured continuously.

The intended destination changed the harvest question

The grapes were deliberately picked for verjus rather than conventional wine. The aim was to retain an early-harvest grape foundation that could be developed into a drink for food, rather than to maximize a single ripeness number.

A bulk verjus sample recorded on 6 September measured 11 °Brix and pH 2.8. In the project’s tasting notes it was balanced and very drinkable. The second verjus harvest on 17 September yielded 11 kg of bunches. Fruit from the later, final harvest on 27 September measured 13.5 °Brix and was spontaneously described as very sweet by several tasters.

These were different samples and harvest stages, not a controlled series of identical berries. The sensory descriptions were not a blinded panel result. Their value was in revealing a question that Brix alone did not settle: how do acids, sugars, aroma and texture combine into perceived flavour?

Extraction became part of the design

The field record compares a high-yield juicer with a manual Italian processing machine that separated seeds. The manual route gave a clearer juice at a recorded yield of approximately 0.6 litres per kilogram, while the juicer gave greater extraction and drier pomace.

This was a practical comparison during the season, not an equipment trial on matched, replicated samples. Even so, it brought the project into food processing: the way fruit is extracted changes the material available for the next decision.

Clarification, judged through the senses

At Aroma Lab, clarification was treated as an experimental variable. Appearance mattered, but a visually clearer liquid would be of little interest if its aroma and flavour became less expressive.

Three PVPP clarification trials, recorded as P25, P50 and P75, were compared with a control. In the project team’s assessment, P75 gave the strongest combined result in colour, aroma and flavour perception. P50 was acceptable but less convincing; P25 brought insufficient improvement.

The most revealing note concerned time. Compared with the control, the preferred sample felt less immediately juice-like. Its sensory onset was more gradual before reaching full fruit intensity. That describes how the sample was experienced, not a laboratory measurement of aroma release.

Subsequent SIHA Puranit trials, B50 and B100, did not noticeably diminish aroma in the bench comparison. B100 gave better clarity. Together, P75 and B100 now form the working Harvest 2026 processing baseline.

These are internal development observations. They establish a useful next step for this material, not a universal optimum, a completed safety assessment or a transferable production recipe.

Letting the grape base speak

The current direction for VITALIJA No. 01 is the pure early-harvest Vitalija/verjus base. Earlier concept renderings carried basil and bitter citrus on the label; those words are not confirmation of ingredients in the final product. The present work asks how much character can be revealed through the fruit, the harvest decision and careful processing.

This is where the engineering and agricultural record meets food chemistry and sensory product design. Measurements help describe the material. Tasting helps decide what to investigate next. Neither replaces the other.

From sensor to flavour

The phrase names a way of working: observe the plant, record the conditions, taste the fruit, make a decision, then test what that decision changes. A digital twin is useful here as a framework for learning, not as a substitute for an attentive grower or a palate.

The research began with one vine. Its credibility rests on keeping that scale visible, acknowledging what remains unknown and allowing the product to change when the evidence changes.

For the broader dining context, read Beyond 0.0. Explore Aroma Lab and the Botanical Cuvée project to see where this work is heading.

Source and scope: Vitalija 2026 · Research Case · Year Zero, and the project owner’s subsequent clarification-trial update. The field study centres on one vine. Bench observations are reported as observations; replicated trials, standardized sensory work and laboratory validation remain distinct requirements.