Tone and safety

Settings

Internal command center for finding high-signal conversations, drafting replies, and publishing only after review.

Live operator workspace, no fake actions in header.

Tone profile

The product needs explicit voice control, not vague prompts.

  • • default tone: founder-led
  • • emoji default: off
  • • service mention: controlled
  • • reply length: short to medium
  • • banned phrases and robotic openings

Publish safeguards

Manual first, with hard rails in the system.

  • • approval required by default
  • • duplicate phrase detection
  • • blocked accounts and blocked phrases
  • • per-day caps
  • • cooldown between replies