AI Moderation (toxicity filter)

A free BEV dashboard feature

Switch it on and set it up on the BEV dashboard. Everything below can be changed there, with a preview before anything is posted.

Settings

When a message is harmful (a choice)
Choices: Tell the staff only, Delete it, Delete it and time out for 1 hour, Delete it and warn them
Look for (several choices)
Choices: Bullying and harassment, Hate speech, Threats and violence, Sexual content, Self-harm (staff are told, the member is sent support), Scams and phishing
How strict (a choice)
Choices: Relaxed: only clear cases, Normal, Strict: borderline too
Check pictures too (on/off)
Costs more: each picture is one extra check
Channels to skip (channels)
Most checks per hour (a number)
Keeps the cost under control. Messages past this aren't checked until the hour is up