This is still an active research area. I know of two main approaches in order to get "on demand" single photon sources.
One of them involves spontaneous parametric down-conversion combined with multiplexing. Temporal multiplexing scales poorly regarding loss, but frequency multiplexing might help. Here people use multiple two-mode squeezers and frequency shift the photons for example with an RF ramp on a phase modulator synchronous with the passing photon. (see e.g. https://arxiv.org/abs/1907.10355 )
The other is quantum dots, which are single emitters, so they are deterministic, as long as one can direct all the emission into a spatial mode of interest. As far as I know the main challenge with them is that their spectrum depends very much on their local environment, so photons from different dots end up being partly distinguishable, unless cherry picking is done. https://en.wikipedia.org/wiki/Quantum_dot_single-photon_source
Please correct me if I got something wrong here.