Just to clarify: A website will always use the user's browser to make a request to a client. This can be confirmed by sniffing things like the request's IP address. Using another computer to make a http call for you would be high dangerous and would likely come into the catagory of Remote Access Tooling, a form of system hacking. Imagine if you could query a bank's website and make the bank think you were hitting it from an internal computer!
It is on these grounds that you are able to add security such as blacklisting IP addresses based on Geography, or altering content (such as Netflix UK vs Netflix US etc).
If you want to implement a secure architecture I would recommend the following (this is what we do in an investment bank)
- Expose a few end points publically, this is what will be used to communicate from hte front end to the back end, and essentially bridge the gap.
- These servers can then make calls to your secured network and can hide any private network traffic.
- Return the data from the private network to the front end via the public end point.
Think of the public endpoints a a cloacking device, as it provides a controlled entry point which can be covered with defences like CORS, Authorisation and Authentication etc and also masks any objects back into public viewmodels.
Let me know if you want more information :)