The objective of this study was to evaluate the effects of oak tannin extract (OTE) added in forage before ensiling on dairy cows fed at 92% of their digestible protein requirements. Six multiparous lactating Holstein cows were used in a crossover design (two treatments × two periods). The control treatment (CON) was based on a diet including 50% of grass silage, whereas the experimental treatment (TAN) included grass silage sprayed with OTE (26 g/kg DM) just before baling. Milk yield (on average 24 kg fat protein corrected milk per day) was not affected, but both milk and rumen fatty acids profiles were impacted by OTE. Nitrogen intake (415 g N per cow per day) and nitrogen use efficiency (NUE; 0.25 on average) were not affected, but a shift from urine (-8% of N intake relatively to control, P = 0.06) to faecal N (+5%; P = 0.004) was observed with the TAN diet (P ≤ 0.05). Nitrogen apparent digestibility was thus reduced for TAN (-3%; P ≤ 0.05). The effect of OTE on ruminal and milk FA profiles suggests an impact on rumen microbiota. Nitrogen isotopic discrimination between animal proteins and diet (Δ15N) was evaluated as a proxy for NUE. While no differences in NUE were observed across diets, a lower Δ15N of plasma proteins was found when comparing TAN v. CON diets. This finding supports the concept that Δ15N would mainly sign the N partitioning at the metabolic level rather than the overall NUE, with the latter also being impacted by digestive processes. Our results agree with a N shift from urine to faeces, and this strategy can thus be adopted to decrease the environmental impact of ruminant protein feeding.