Abstract. The oxygen (𝛥17O) and nitrogen (𝛿15N) isotopic compositions of atmospheric nitrate (NO3-) are widely used as tracers of its formation pathways, precursor (nitrogen oxides NOx = nitric oxide NO + nitrogen NO2) emission sources, and physico-chemical processing. However, the critical lack of observations on the multi-isotopic composition of NO2 maintains significant uncertainties regarding the links between the isotopic composition of NOx and NO3-, which may bias estimates of the NO3- formation processes and the distribution of sources. We report here on the first simultaneous atmospheric observations of 𝛥17O and 𝛿15N in NO2 and NO3-. The measurements were carried out at sub-daily (ca. 3 h) resolution over two non-consecutive days in an Alpine city in February 2021. Important diurnal variabilities are observed in both NO2 and NO3- multi-isotopic composition. 𝛥17O of NO2 and NO3- range from 19.6 to 40.8 ‰ and 18.7 to 26 ‰, respectively. During both daytime and nighttime, the variability of 𝛥17O(NO2) is mainly driven by the oxidation of NO by ozone, with a substantial contribution from peroxy radicals in the morning. NO3- local mass balance equations, constrained by observed 𝛥17O(NO2), suggest that during the first day of sampling NO3- was formed locally from the oxidation of NO2 by hydroxyl radicals during the day, and via heterogeneous hydrolysis of dinitrogen pentoxide during the night. For the second day, calculated and observed 𝛥17O(NO3-) do not match, particularly daytime values. The effects on 𝛥17O(NO3-) of a Saharan dust event that occurred during the second day and winter boundary layer dynamics are discussed. 𝛿15N of NO2 and NO3- ranged from -10.0 to 19.7 ‰ and -4.2 to 14.8 ‰, respectively. Consistent with theoretical predictions of N isotope fractionation, the important variability of 𝛿15N(NO2) is explained by significant post-emission equilibrium N fractionation. After accounting for this effect, vehicle exhaust is found to be the primary source of NOx emissions at the sampling site. 𝛿15N(NO3-) is closely linked to 𝛿15N(NO2) variability, which bring further evidence of fast and local processing, but uncertainties on current N fractionation factors during NO2 to NO3- conversion are underscored. Overall, this detailed investigation highlights the potential and the necessity to use 𝛥17O and 𝛿15N in NO2 and NO3- to trace quantitatively the sources and formation chemistry of NO3-, particularly in urban environments in winter.