Show HN: AmiBot – A Slack bot to detect toxic messages

Hi everyone! This is a quick little Slack bot I've written using Google's new PerspectiveAPI.

Essentially she will analyse all the messages in any channel you invite her to, if a team member breaches the toxicity threshold she will inform them via a DM.

She works on a 5 minute buffer so it should only trigger if someone is being particularly spiteful.

None of the messages are logged or stored by myself or PerspectiveAPI.

In future I hope to extend it to help admins of large slack teams moderate members by informing them of offenders and providing more customisation options

What is considered toxic?