What components to use for multi-threaded HTML parsing on VC++ using proxies?

Question

HTML

Statement of the problem : it is necessary to parse a large number of websites on a daily basis (> 100 sites, > 1000 pages) and extract information about products from them. Let's say online shopping. We need multi-threaded work using a proxy (one page (not a site) - one proxy).
Actually the question - please advise a complete "binding" of the final solution with a focus on:

1. Performance and economy (to run multiple threads on the same machine/link).
   
2. Stability and compatibility (pagination, frames, unicode and all other layout features).
   
3. Security (for example, destructive code / virus is transmitted along with the page).
   

What components to use to work with HTML? CsQuery/HtmlAgilityPack... How to access sites through pre-purchased proxies? Do the components work with HTML proxies, or do you need an additional "shim" to use proxies?
I would be wildly grateful for a detailed description and sequence of actions (I'm not a professional programmer).
PL / development environment - VC ++ 2015. I understand that this may not be the best PL for solving such problems. But I ask you not to raise the issue of choosing / changing the YP. Interested only in VC++.

Answer 1

Processed 1k pages in Python 2.7, used from multiprocessing import Pool. Look at my farrows, there were links somewhere.