Skip to content

Captions and voice

Each scene has one caption field. It is spoken as the voiceover and, when Show captions is on, shown on screen as the subtitle. Voiceover has a voice, a style preset, and the caption as its text.

Inside a caption you can insert inline tags such as [whispers], [laughs] or [sighs] to colour a moment. A tag button appears near your selection with the curated list. Use at most a couple per line.

  • Tags are spoken as delivery, never read aloud as words.
  • Tags are stripped from the on-screen caption at render. In the editor the caption shows exactly what you typed.
  • Tags do not count against the spoken-length limit.
  • 200 spoken characters per scene, tags excluded. A voiceover longer than its scene is cut off at the scene boundary — match durations to audio (Manan does this with fit duration to voiceover after a batch finishes).
  • Prompt fields (image, motion, music) hold up to 2,000 characters.
  • @-mentions are disabled in the voiceover field, because the text goes to the voice engine verbatim and a mention would be read aloud.

The voice catalogue lists prebuilt voices with their character and the style presets. Keep one narrator voice across a project unless scenes clearly speak as different characters. A standing voice choice belongs in a project instructions document so Manan uses it without asking.