The Savitzky-Golay filter is explained here. It looks to me that any Savitzky-Golay filter can equally be considered an FIR filter. Am I right about that?
Yes. The Savitzky-Golay filter is a weighted sum of a finite number of data samples. So it is indeed an FIR filter (albeit, not causal, because it uses samples from the past and the future). And it is indeed a low-pass filter (smoother) as seen from the frequency response of the filter on the same page.