508552 URL Page Scraper

I already have an application like this (someone here on SL developed one for me a few months ago) but... as it's rather slow and unstable, I'd like someone to develope a better one. The project description, is actually quite simple. I don't need any fancy visual design... just a basic, simple user interface: a browser window, with a few buttons like "Pause" "Start", "New Search", etc. I will go to a URL, using this application. That URL will be an Amazon page ([login to view URL], or [login to view URL], or other...). I will perform a search for a product (example, a book). My search however, will be very general (returning many results). For example... maybe 7000 results based on my search. I might filter the search down a bit, to show less results, or, I might not. Regardless, when I hit "Start", the application will then scrape details about each of those items found at this point (so let's say if my search to this moment, led me to 6402 results, then... when I hit start at this point, the script will crawl through each subsequent page in the Amazon search results, and collect the data I need about each product, and populate this data into a csv file (the data I need, I will outline later, but it will always be the same data- and is easily available via Amazon's API). Again, I already have an app that does exactly this, but... it's a bit slow (only secures about 5 items per second) and sometimes stops in the middle of a session... meaning, I have to re-do the entire process from scratch. Note: The output, will actually be generated into three different files (all will have the same data, but... I just need three different file types/templates generated, for different uses with this data). They will all be in csv format, so it'll just be a mater of telling the script which data to put into which column in each file. Last... Amazon search results, are limited to 4000 items (if you put in a search for the world "Good" in books, you might get a result of 50,000 books that have this search key in them but if you try to browse page by page, you'll be stopped at 4,000. The app will have to be able to secure data for all results, even if the url (start point) will show more than 4000 results. There is a way around this, and I will explain it to the winning bidder. This basically will require the following skills: a good understanding of Amazon API, some (entry level) skill with database population, and... absolutely no design skills (again - I dont care about design- just functionality).