Cool idea, but no Linux support and only Twitter and Reddit? Does it work by simulating user activity? I'd be worried about anything using the API that would trigger some internal alert/archival due to thinking it is a bot trying to remove evidence. I think the biggest thing to keep people safe is not to post under their real identity, and try to use a service (VPN, proxy, etc) that doesn't trace back to you. Web browsers are pretty guilty here in terms of what they permit to be tracked by default using JS that can almost immediately identify a user based on browser version/OS/etc.