Ok, great! I will be waiting for an update then ;o)

This is how the "stopping" looks line in Messenger. The bot simply does not reply. It is stuck! Note that all entered keywords are listed in the node: "pareizi, ja, jaa, taisniba, taisniiba, jā, taisnība". If I enter something that is not in the keyword list, the bot moves to the No Match Node:

So the issue is with your platform not being able to reply when triggering keywords in the particular node.

Hi, I see that you have misunderstood the issue with keywords. There are multiple actually:

1) I belive keywords with diacritics marks often do not trigger at all (seems like some component - probably the one that you use for tokenisation or the "NLP" stuff - does not support UTF-8).

2) Sometimes for some reason keywords without diacritics marks are not triggered when using the bot through Messenger or Skype.

I did not by no means try to capture a word without diacritics marks as a keyword with diacritics marks (Although I know that some search engines also do this as a Romanisation feature - I do not think that it is smart to have such a feature for keyword processing ... especially because words with diacritics marks can mean totally different things ... as I explained above).

... These two actually correspond to the first and second example. If you read my comment (above) carefully (without subjective pre-conditioned bias), you will understand that I explained these two issues (and the examples clearly show the issues).