I looked at the gothub, I'm not fully sure their approach but if they're just scraping the data you hit rate limits pretty fast and you'd have to use proxies. They also don't seem to have full feature parity with Github which is our end goal.

In terms of self-hosting we will add instruction but our vision is mostly people using our hosted version. A lot of the optimization comes from the caching.