c++ website scrapper to be use from php script

I currently do all that with php scripts and Curl but as you know php is not multithread so i need something faster for my website. I beileive that is a small program so price must reflect that I am not doing it by myself because i can't program in c++ So if you think this is hard to do or long just pass you will save me some time and effort ! I need a c++ program that will run on my debian lenny server, the program will be called from a php script. Program have for principal function to connect to websites and save the source page to a file. No interface needed Anyway the program must be multithread and run a minimum of 20-50 url in same [login to view URL] must execute standard redirection header and following. Program can be also be called alot of time in same time so make someting that can work like mysql many connection in same time. When program finish the job i need him to return a success state so i can continue my code in my script. You can use libcurl if this work with the program . You can also make a php module if you whant i just need something fast and easy to install. All sources code need to be sent so i can compile it on my own server (send also compiled project) Program must take no more that the longest page load to execute so if you connect to 30 urls the longest of these 30 pages will be the execution time. Read and save file is very fast so i do not count time on that ... But i do not want my server to crash so advice me on how many thread in same time i can run if you said 10 000 make sure the program can handle them Check files that are inclued you will find alot of informations