Skip to content
Clevera Docs

Use voice tags in narration

Add pauses, laughs, sighs and tone cues to a narration line with voice tags. Every tag, what Built-in, ElevenLabs and avatar voices do with each one, and tips for natural results.

Updated 5 min read

Voice tags are short cues in a narration line, such as a pause or a laugh, that change how the AI voice reads it. This article lists every tag, what each kind of voice does with it, and how to get good results.

Before you start

  • Voice tags work in the narration of how-to videos and product demos: the lines in the Script tab (see Edit the script and voice-over) and the voiceover of a List scene.
  • A tag takes effect when the line is recorded, so saving a line with a new tag uses credits like any other re-recorded line.

Add a voice tag

  1. Open the line

    In the Script tab, hover a step and click Edit narration.

  2. Put the cursor where the cue goes

    Click in the text right where the pause or sound should happen, for example after "Settings,".

  3. Pick a tag

    Type /. A list opens under the line ("Voice tags · adapted to each voice"). Keep typing to filter it, use the arrow keys, and press Enter, or click a tag. You can also click Voice tag under the text ("Add a voice tag (or type /)"), which opens a searchable Voice tags picker ("Each voice engine reads these its own way.").

    The Voice tags list open under a narration line after typing a slash, with Short Pause highlighted.
    Type / where the cue should go.

  4. Save

    Click Save or press ⌘ Enter / Ctrl Enter. The line is recorded again with the tag.

The tag is written into the text in square brackets, for example [short pause]. You can also type it by hand; the brackets and the exact words matter, capital letters don't. In the Script tab a saved tag shows as a colored chip with its name; on the timeline you see the bracketed text.

A narration line in the Script tab with a Short Pause tag between two phrases.

Every tag

Tag Written as What it asks for
Short Pause [short pause] "Brief pause (~250ms)."
Medium Pause [medium pause] "Standard pause (~500ms)."
Long Pause [long pause] "Significant pause (~1000ms+)."
Laughing [laughing] "Inserts a laugh sound."
Uhm [uhm] "Inserts a hesitation sound."
Sigh [sigh] "Inserts a sigh sound."
Sarcasm [sarcasm] "Imparts a sarcastic tone."
Shouting [shouting] "Increases the volume."
Fast [extremely fast] "Increases the speed."
Curious [curious] "Spoken word with curious tone."
Bored [bored] "Spoken word with bored tone."

How each voice reads them

Different voices understand different cues. Clevera translates each tag for the voice that records the line, and drops the ones a voice can't perform, so a tag is never read out loud as a word.

Tag Built-in voices ElevenLabs voices Avatar voices
Short Pause Short pause Short pause Ignored
Medium Pause Short pause (there's no medium step) Pause Ignored
Long Pause Long pause Long pause Ignored
Laughing Laugh Laughs Ignored
Uhm Ignored Says "uhm" Ignored
Sigh Sigh Sighs Ignored
Sarcasm Ignored Sarcastic delivery Ignored
Shouting Shout Shouting Ignored
Fast Ignored Speaks quickly Ignored
Curious Ignored Curious tone Ignored
Bored Ignored Bored tone Ignored

The column that applies is the source of the video's current voice, shown as a tab in the voice picker. See Change voice, music and sound.

Built-in voices use a pause, laugh, sigh or shout as a short event at that point. The tone tags (Sarcasm, Curious, Bored, Fast) have no effect on them. Uhm is left out for every language because a filler sound would only suit English.

ElevenLabs voices from your own account understand all eleven tags; how strongly a voice reacts depends on the voice.

Avatar voices read the words only. Every tag is removed before recording.

Tags and the rest of your video

Question Answer
Do tags show in captions? No. Captions are made from what the voice actually says, so a tag never appears as a word.
Do tags survive a translation? Yes. The translation keeps your tags where they were in the sentence, and the new language's voice reads them as in the table above.
Do tags carry over when I change the voice? Yes. The text keeps its tags, and each line is recorded again with the new voice's reading of them. A tag the new voice can't perform is skipped.
Does a pause make the step longer? Yes. Pauses add to the line's length, and a step lasts at least as long as its narration.
Can I write my own tag, like [whisper]? Only the eleven tags in the list are treated as tags. Other text in square brackets is sent to the voice as written and may be read out.

Tips

  • Use pauses for structure: a Short Pause after a menu name, a Long Pause before a new idea. They're the tags every voice except avatar voices respects.
  • Put a tone tag right before the words it should color: "That's it. [curious] Want to see what happens next?"
  • One or two tags per line is plenty. Many tags make a line sound staged and harder to time.
  • If a tag doesn't seem to do anything, check the table above: the current voice may not support it.
  • Press Regenerate voice to hear another take. Each take can sound slightly different.
Edit the script and voice-overWrite and re-record narration lines.Change voice, music and soundPick the voice that reads your tags.Use your own ElevenLabs voicesThe voices that read every tag.Use sync and pacingMatch the footage to the new timing.