Feature Request: TTS support with different voices

[lopresti851]

Are there any plans for some kind of built-in TTS support, especially with different voices for characters and narration. Right now, if you want the story read aloud, you have to use a separate TTS app or browser extension, and everything comes out in the same voice. It’d be amazing if Waidrin could either natively handle TTS or make it easier to send lines to different voices automatically.

Is that kind of feature on your radar, or do you know of any workarounds/plugins that could help with this right now?

[lee-b]

FYI, some of the tts models support multiple voices. I think Orpheus is one such model, for example. Ideally Waidrun would:

1. Support various TTS engines, with a cheetah-like template specification for the expected output format, including some variable for the voice/speaker_id to use

2. Within waidrin, track which attributes speakers have (e.g., is_narrator, is_pc, is_companion, is_npc, name, sex, age, character_id), and you'd be able to map one ore more of those attributes to variables that get injected into the tts template format as it's being expanded and send to the TTS engine.