Skip to main content
Inline tags control delivery at the token level. Insert them anywhere in input, and the model adjusts the surrounding speech. For example:
Tags fall into five categories:
  • emotion: <|emotion:…|>, such as elation, fear, anger.
  • style: <|style:…|>, such as shouting, whispering.
  • sound effects: <|sfx:…|>, such as cough, sneeze.
  • prosody: <|prosody:…|>, including speed, pause, pitch, and expressiveness.
  • other: <|reset|>, which resets delivery to neutral.
Only the tags on this page are accepted. Any other <|...|> in the input is refused with 400 invalid_field_value, naming the tag, so a typo fails instead of being read aloud. Recommended usage:
  • Set delivery with delivery tokens. Emotion, style, speed, pitch, and expressiveness tags set how the text after them is delivered. Place them at the start of the input to set delivery for the whole turn. To change delivery partway through, add a new group of delivery tokens where the change begins. A new group replaces the previous one entirely: in <|emotion:anger|><|prosody:speed_fast|>Text one. <|emotion:sadness|>Text two., “Text two.” is sad at normal speed. Use <|reset|> to clear all delivery tokens and return to neutral delivery.
  • Place positional tokens exactly where they take effect. <|prosody:pause|> and <|prosody:long_pause|> go where the break should fall, and each <|sfx:…|> goes right before the sound it triggers.
  • Pair every sound effect with onomatopoeia. A <|sfx:…|> token works best when the matching written sound follows immediately, such as <|sfx:laughter|>Haha, <|sfx:sigh|>Uh, or <|sfx:sneeze|>Achoo. The written cue helps the model realize the sound effect.

Emotion

Style

Sound effects

Sound effects are vocalized — produced in the speaker’s voice — not mixed-in audio assets.

Prosody

Prosody tags control speed, pitch, pauses, and overall expressiveness.

Other