Granted I was using IPv4, but I've had a lot of trouble using Wireguard to my Linode server for casual browsing. Google became unusable with repeated captchas, several gaming apps blocked my login. I wish web owners wouldn't automatically classify datacenter IPs as malicious. There must be better ways to block spam/bad actors than by the classification of their source IP.
In American English and the Queen's English, "I don't have IPv6, how to access my node?" is bad grammar and should be "I don't have IPv6, how do I access my node?". As a German selling things in Euro, you are clearly using International English and are free to make your own rules - (I do see this sentence structure a lot with non-native speakers). If that is the standard, us Americans will happily mind our own business. Not sure about the Brits.
Github works just fine, all VMs come preconfigured with a NAT64/ DNS64 nameserver server so you can reach IPv4 only systems as long as you try to resolve them via a domain.
The idea is to use the built-in reverse proxy for proxying HTTP(S) requests and provide an IPv4 + IPv6 endpoint for your domain.
That means your end users won't even notice the server is running on IPv6 only
Google has some interesting stats on the IPv6 adoption rate: https://www.google.de/ipv6/statistics.html