First a url is submitted to the database via the form on our main website, through an image tag from another webpage, or through a bookmarklet.

Next, a cluster of screenshot generation servers running a configuration of apache/mysql/php and other custom software sits and waits for new jobs to appear in the database. Once it gets a job, it uses a Internet Explorer to connect to the website, wait for the site to load then take a screenshot and save the image. While on the site, it also grabs useful information about the page such as its title and meta tags (to be used later).

Lots of error checking is done in case the site does not respond. If there is an error, detailed error analysis logs get put into the database so we can investigate the problem later.

Because of the distributed nature of the screenshot generation process, the website can scale to take 10 shots or 10,000 shots by adding more generation machines. The generation machines can also be run in a virtual machine process so they can be cloned and deployed quickly, and it also provides an easy way to recover in case a malicious script sneaks it's way into a machine. We have done a lot of work to prevent the execution of harmful scripts on our generation machines.

As screenshots are being generated our website uses AJAX to keep visitors and administrators updated about the progress. Administrators have a comprehensive stats interface that not only shows all current and past jobs and their status, but also gives valuable information such as error percentage, average number of screenshots being generated per minute, and the time elapsed to take each screenshot. While visitors are waiting for their screenshot to be generated they get to watch a scaled down version of the admin section similar to digg spy which constantly scrolls the latest screenshots. This spy interface also pulls relevant information obtained from the site such as it's title tag, meta tags, and keywords.

To sign up for our members only area click here..