In the scope of wireguard it’ll just be a matter of you building appropriate firewall rules.

Since you want their internet traffic to go through you then i assime you’re effectively pushing a 0.0.0.0/0 route to your clients. You then need to add firewall rules on your server to block traffic to its local subnet and in the future allow traffic to only your jellyfin server.

This is also pretty simple and nothing wrong with that setup.

You did not answer what VPN tech you are using.

Without that knowledge i would recommend setting up tailscale and having your users use that. If you want to be fully self hosted you can also run Headscale as the control plane instead of relying on Tailscales own service.

I recommend tailscale as it is very easy to grant a user privileges to ONLY use an endpoint as an exit node but also grant access to any other endpoints as needed (such as your future jellyfin server) via theor ACLs.

Best practices comes down to what you do or do not want the VPN clients to access. This mostly comes down to routing and firewall rules.

So, what should your users have access to?

Also what is the vpn?

I’m not entirely sure what the actual question is. Can you rephrase what exactly you are trying to accomplish?

Error message? Nextcloud logs?

Can’t tell you whats happening without information about what’s happening other than “it doesn’t work”.

Why not just run a reverse proxy container on the server hosting the rest?

As others have said, cool concept, awful name.

Bad name aside Windows only client support is a big letdown and makes the application useless to me.