Woebot seems interesting but the fact that it's an app on a phone makes me immediately suspicious. Why not just have a website, unless the intent is data-mining of some sort? And calling it a "university research project" is underselling it a bit considering they've received series A funding and at this point should be considered a full-fledged member of the silicon valley startup machine.
It could well be effective and made with benign intentions, and I have no doubt that its results are better than traditional shrinks since the bar to improve on that is pretty low. But maybe it's not just my cup of tea since I don't see how what amounts to ELIZA 2.0 can provide meaningful engagement and debate (and the limitations of having to type on a phone definitely dissuade from from detailed conversation). And after a while the repetitiveness of hearing the word "mindfulness" would drive be mad (assuming that the bot does this; I have not used it aside from looking at screenshots of transcripts). Overall I think the anonynimity of imageboards is better in terms of providing meaningful, "stress-lowering" interaction.