Should have been more explicit with my question., I'll try to do it now.
With tools I meant, which standard libraries language serverside, training corpus etc., I hope to have made clear.
Hmm, nothing standard, really... I just used vanilla python and the corpus is some texts and ungrammatical comments I've assembled from various sources...