The radio detection technique of cosmic ray air showers has gained renewed interest in the last two decades. While the radio experiments are very cost-effective to deploy, the Monte-Carlo simulations required to analyse the data are computationally expensive. Here we present a proof of concept for a novel way to synthesise the radio emission from extensive air showers in simulations. It is a hybrid approach which uses a single microscopic Monte-Carlo simulation, called the origin shower, to generate the radio emission from a target shower with a different longitudinal evolution, primary particle type and energy. The method employs semi-analytical relations which only depend on the shower parameters to transform the radio signals in the simulated antennas. We apply this method to vertical air showers with energies ranging from $10^{17}$ eV to $10^{19}$ eV and compare the results with CoREAS simulations in two frequency bands, namely the broad [20, 500] MHz band and a more narrow one at [30, 80] MHz. We gauge the synthesis quality using the maximal amplitude and energy fluence contained in the signal. We observe that the quality depends primarily on the difference in $X_{\text{max}}$ between the origin and target shower. After applying a linear bias correction, we find that for a shift in $X_{\text{max}}$ of less than 150 $\text{g}/\text{cm}^2$ , template synthesis has a bias of less than 2% and a scatter up to 6%, both in amplitude, on the broad frequency range. On the restricted [30, 80] MHz range the bias is similar, but the spread on amplitude drops down to 3%. These fluctuations are on the same level as the intrinsic scatter we observe in Monte-Carlo ensembles. We therefore surmise the observed scatter in amplitude to originate from intrinsic shower fluctuations we do not explicitly account for in template synthesis.