Google Keyword Tool Scraper(repost)

We need a application/bot made that should return results from cycled through keywords from the Google Keyword Site - [login to view URL] I will define the cycle required below: 1. Keyword entered into 'word or phrase' text box on website. 2. Ensure that the 'Match Type' - '[Exact]' tick box is selected Only. 3. Click the Search button. 4. Results are displayed. At the moment the competion bar on the results is reported as either low competition, medium competition and high competition but as you can see the bar varies in terms of how much green colour fill it has. If possible can we translate this into a percentage and return the percentage figure for our results? 5. The following statistics - the keyword, competition percentage, global monthly searces and local monthly searches for each keyword under the 'Search Terms' and 'Keyword Ideas' sections need to be added to a database/spreadsheet that is constantly updated and viewable and sortable in desc/aesc order. 6. The program then needs to make sure it remembers the keywords searched and take a not already searched keyword from the 'keyword Ideas' section on the page and search that as the main keyword from point 1 onwards and the cycle continues! The idea is that with this cycle we'll be able to search through the thousands of search terms that exist in the Google Keyword Database and return the results. When storing the keyword data we need to be able to sort the data as we please through the adjustment of the parameters - competition percentage, global monthly searches and local monthly searches as we deem necessary. So if i wanted to search our database for only keywords with global monthly searches => 40,000 i should be able to. Also i want to be able to start and pause this application where it left off whenever i need to, as you can imagine there are millions are keywords so to capture all of them in a couple of days will be impossible so i plan on running this over a long perioud of time. I think we would need to use proxies for this which i can supply with as im not sure scraping only with my host IP would work for long. If there are any suggestions that you have to make this better, then please do let me know too :-)