We consider a consistent vehicle routing problem for the delivery of parcels with electric vehicles. Stemming from a real-world problem, we assume that vehicles can only be charged with electricity between their delivery tours in the morning and their pickup tours in the afternoon. For this purpose, a charging station with a limited amount of charging slots is available at the depot. We aim at generating a set of vehicle routes that are driver- and time-consistent and efficiently use limited charging resources, while optimizing the sum of vehicle fixed cost, vehicle/driver operating time, arrival time consistency and driver consistency. We present a mathematical model to describe the problem in detail. For solving the real-world problem, a template-based Adaptive Large Neighborhood Search is developed, complemented with constraint programming for charging management and quadratic programming for delivery and pickup trip scheduling. Computational experiments for different settings and scenarios, based on data from an Austrian parcel delivery company, are presented and analysed.