Skip to main content

Detect toxic language in Telegram messages

This workflow detects toxic language (such as profanity, insults, threats) in messages sent via Telegram. This blog tutorial explains how to configure the workflow nodes step-by-step.

workflow-screenshot

Telegram Trigger: triggers the workflow when a new message is sent in a Telegram chat.

Google Perspective: analyzes the text of the message and returns a probability value between 0 and 1 of how likely it is that the content is toxic.

IF: filters messages with a toxic probability value above 0.7.

Telegram: sends a message in the chat with the text "I don't tolerate toxic language" if the probability value is above 0.7.

NoOp: takes no action if the probability value is below 0.7.