Rolling predict without refit?

[EgorKraevTransferwise]

I need to do a rolling prediction, that is, for t = T0, T0+1, ..., T1 I need the model to accept the time series up to t and predict the next, say, 5 steps each time.

In the Orbit models, is it possible to do this without having to refit the model on each new time step (the computational load and speed are prohibitive otherwise).

So fit the model on df[:T0], then use the fitted model to predict df[t:t+5] given df[:t], for each t, without refitting? Or is refitting an integral part of predict, like in Prophet, where you have a nonlinear decomposition into 'trend', 'seasonal' etc components, which has to be redone each time we change the past dataset?

Is your feature request related to a problem? Please describe.
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]

Describe the solution you'd like
A clear and concise description of what you want to happen.

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
Add any other context or screenshots about the feature request here.