Voices
The 127 reading voices available, with the names and style numbers you type to choose one.
Speech synthesis uses VOICEVOX. Each speaker has styles — Normal, Sweet and so on — and the voice changes with the style. The number after a style name is its id.
To change voices, send !speaker <name> in the web chat, or !speakers to pick from a list. Numbers work too. Your own voice and the Space default are separate settings, and yours wins where both are set. See the command reference.
Choosing one
For a long Space, leave the default in place and send a short message to hear how it lands. If something is off, adjust the speed and volume before changing voices — speed alters the impression more than people expect.
Giving each participant a different voice means you can tell who wrote a comment by ear. A calm default for the Space, with listeners picking their own, avoids the two clashing.
Reading voices and the singers used by !sing are separate lists; check singer numbers with !singers. The singing styles below cannot read.
The synthesiser is Japanese. English text is pronounced with Japanese phonetics — passable for a few words, awkward for whole sentences. There is no English voice.