1Text to Speech
Speak text in a chosen voice. Takes its text from an upstream step — a translation, for instance — or from text typed at run time.
3 inputs asked at run
2Approval
Wait for someone to approve before the flow continues.
Nothing asked at run
3Quality Control
Check an image, audio, or video for compliance and cultural fit, and get an Approved, Review, or Blocked verdict.
4 inputs asked at run
4Switch
Send the flow down a different branch depending on a value.
Nothing asked at run
5Approval
Wait for someone to approve before the flow continues.
Nothing asked at run