it'd be cool if it were possible but, this is is from the manual "The 902 examines the differences in dB between the high frequency and full-bandwidth portions of the signal, allowing de-essing of signals which change in level by as much as 60 dB. Conventional de-essers require readjustment of their threshold control when a vocalist drops from singing voice to a whisper- ing voice. By contrast, the 902 does not even have a threshold control to require adjustment"
before i saw that i thought it would be maybe possible but just very tricky and also with unwieldy results. you can't have nebula set up to do anything like that, afaik. this deeser looks more advanced than the basic deeser and i don't see nebula having the ability to do that, at all. splitting a signal into two bands and calculating level of one to match the other... no. not possible.
but just looking at a basic deeser, even a most basic design would be tricky for nebula to do. nebs internal filter for the internal/external sidechain only allows high passing, for one thing. the filter shape of the hardware deeser would be pretty important to have somewhat accurately recreated in nebula for the result to really be similar. if you could sample that, then have a program for it on the ext sidechain, that could work, but then the compression aspect of the deeser still has to be set up to be similar.