Pyrolysis is a key process in all stages of wood burning from ignition to extinction. Understanding each stage is crucial to tackle wildfires and assess the fire safety of timber buildings. A model of appropriate complexity of wood pyrolysis and oxidation is missing, which limits the understanding of fires fuelled by wood. Progress towards this aim has been slow in recent years, as the role of chemical kinetics is still debated. Three predominant theories hypothesis that chemistry is either infinitely fast (de Ris), a function of char depth (Atreya), or a function of heat flux (Suuberg). This paper proposes a novel multi-scale model of wood pyrolysis and oxidation for predicting the charring of timber. The chemical kinetics sub-model was previously validated at the microscale (mg-samples). We favourably compare the complete model against a large range of mesoscale experiments (g-samples) found in the literature of different moisture contents (0-30%), heat fluxes (0-60 kW/m2), oxygen concentrations (0-21%), grain directions (parallel/perpendicular), and combinations thereof. The model was then used to calculate the transient Damköhler number of wood at different depths and heat fluxes. This analysis showed that chemistry and heat transfer are both important at all heat fluxes and stages of burning relevant to fire, which unifies the three theories by Suuberg, Atreya, and de Ris. We argue that the model is of currently appropriate complexity to predict the charring of timber. These findings improve our understanding of wood pyrolysis and the modelling of timber burning across scales.