We develop a bottom-up model of inflation in the euro area based on a set of augmented Phillips curves for seven subcomponents of core inflation and auxiliary regressions for non-core items. We use the model’s disaggregated structure to explore which factors drove the deterioration in forecasting performance during the pandemic period and use these insights to improve on the ability to forecast inflation. In the baseline, the projection for core inflation is centered above 3 percent at end-2023, while headline inflation is expected to drop quite sharply over 2023, with energy base effects pulling inflation down from the currently very elevated levels to below 3 percent by 2023q4. The confidence intervals around these projections are wide given elevated uncertainty. We argue that the bottom-up approach offers a useful complement to the forecasters toolbox – even in the current uncertain environment - by improving forecast accuracy, shedding additional light on the drivers of inflation and providing a framework in which to apply ex post judgement in a structured way.