Skip to main content
Preferences is the last page of the Voice Engine. It controls how your character behaves in Discord voice channels, how it handles incoming audio, and how its voice messages look in text channels. To open it, go to your character’s sidebar and, under Universal, select Voice Engine → Preferences. The page appears once your character has a voice. Each switch saves as soon as you flip it. Settings that can make your character seem broken show a warning in their info tooltip, and two of them ask you to confirm before they turn on.

Channel Behavior

How your character acts in voice channels. When Allow Interruption is on, you can also turn on Auto-Disable for Busy Channels. It switches interruption off automatically when the number of people in the channel reaches the User Threshold you choose. Turn on Count Bots as Users to include other bots, like music bots, in that count.

Audio Processing

How your character filters what it hears. When Ignore Background Noise is on, two more settings appear:
  • Sensitivity sets how hard noise is filtered: Normal, Aggressive or Very Aggressive. Normal or Aggressive work best. Very Aggressive can filter out normal talking voices.
  • Min Speech Duration sets how long someone must speak before it counts as speech: 100ms (Fast), 250ms (Balanced), 500ms (Strict) or 750ms (Very Strict). 250ms is recommended. The strict options make your character ignore short answers like “yes”, “no” or “hey”.
If your character ignores people, turn Ignore Background Noise off or try Smart Voice Detection instead.

Natural Speech

Enable Natural Speech expands common slang and abbreviations into full words so your character pronounces them naturally. Under Normalization Exclusions, list words that should keep their original spelling and pronunciation, such as gaming terms or your server’s acronyms. Put one word per line or separate them with commas, then click Save exclusions. The counter shows how many exclusions are saved.

SoundBoards

Enable SoundBoards lets your character use your Discord server’s soundboard while it’s in a voice channel. It can see every soundboard sound in that server, pick one that fits the moment, and play it alongside its reply. There’s nothing else to set up. Discord’s soundboard permissions still apply, and your character can only play sounds in the channel it’s in. To guide when it plays sounds, write Soundboard Instructions.

Smart Voice Detection (Beta)

Smart Voice Detection replaces standard noise filtering with AI. Instead of standard voice activity detection, it looks at what was transcribed, decides whether it’s real speech, and decides whether your character should respond. It’s built for cheap or noisy microphones, where Ignore Background Noise struggles.
1

Turn it on

Switch on Enable Smart Detection.
2

Add guidance (optional)

Under Custom Instruction, describe your environment or what to ignore. For example: Ignore keyboard typing and short coughs. Only respond to full sentences.
3

Save

Click Save.

Memory

Memories in VCs lets your character use its Memory Engine in voice channels. It recalls what it knows about people and remembers facts shared in voice conversations, with little effect on reply speed.

Voice Delivery

Native Voice Messages sends voice replies in text channels as Discord voice messages, with a playable waveform, instead of audio links. This doesn’t affect voice channels, where your character always speaks live. Discord can’t attach voice messages to replies to slash commands, buttons or context menus. Those replies use an audio link instead. How often your character sends voice messages is set on Frequency.
  • Allow Interruption: off
  • Smart Voice Detection: on
  • Ignore Bots: on
  • Post Transcripts in VC Chat: on, so people without audio can follow