Making headless screengrabs without PhantomJS

Ok this can be done with PhantomJS and it is probably better/simpler/etc. But if you, for any reason, can’t use PhantomJS to make screengrabs, here’s an over complicated alternative we came up with.

wkhtmltopdf is an easy way of making screengrabs of webpage:
$ wkhtmltopdf “” googlescreendump.pdf
Most package managers (at least Rpm and Aptitude) has it, and it just works. Some arguments may or may not work, depending on version number, and this example is based on the 0.9.9 which Aptitude offers right now.

While this works fine on your local machine, it won’t on a headless server, so we install Xvfb which acts a virtual X-server.
$ xvfb-run –server-args=”-screen 0, 1024x768x24″ wkhtmltopdf googlescreendump.pdf

This will give you a pdf, probably with the webpage not filling the entire page. To get a usable image file, you need to run it through some Image Magick:
$ convert googlescreengrab.pdf -trim -gravity southeast -background none -splice 50×150 googlescreengrab.png
The arguments need to be tweaked for your setup, so you get the entire screengrab and nothing else.

In total, this is what the command looks like:
xvfb-run –server-args=”-screen 0, 1024x768x24″ wkhtmltopdf -s A1 -B 0 -L 0 -R 0 -T 0 –redirect-delay 3000 “” googlescreengrab.pdf && convert googlescreengrab.pdf -trim -gravity southeast -background none -splice 50×150 googlescreengrab.png

So in short. Run PhantomJS if you can.
If not, this will make your client happy.
And make you feel dirty.