Indeed as you mention, complex apps like Taiga are quite time-consuming to package in a way to ensure proper updates for the future. In fact I am right now in the process to iron out the update to Taiga v6 and that price-tag allows us to focus on such things so our users get a smooth experience without missing out on updates (disclaimer, I am one of the Cloudron founders :D )
For sake of our vision we surely would like to make it more cost-effective in the long run, however we are bootstrapped and thus walk a thin line with a focus more on long-term sustainability not just blind growth. (10x cheaper though would realistically even pose an accounting problem with micro-payments or plain transaction fees taking large chunks)
Cloudron also has a full email stack ready to go, however you are totally right about the residential IPs. They are basically all blacklisted by default, which is also why we had to add easy relay-provider support.
This looks really nice and the online demo is quite slick also on Firefox.
I saw that there is either a Qt or GL backend, which made me wonder how this is used then in the online demo or is the browser backend a third, undocumented one?
Also how is input (pointer, key) supported across the already apparently supported platforms?
While I tend to agree from a personal perspective, however the Wekan project especially seems to be able to pump out lots of very nice features at a high rate, thanks to meteor. Additionally to this, we at Cloudron update the app package about twice a week due to that and so far had very little breakage or regressions, compared to other apps. We do not update meteor often though, which is maybe contributing to the stability.