Experiment with client side extraction

[olaven]

Many articles are behind paywalls.
Such articles will be blocked by for the server when it tries to convert.
Extracting text client side would get past this issue if the user
has acces to the articl to begin with.