Remove Eleven v4 performed cues ("[laughs]", "[whispers]", "[short pause]")
from a call transcript line (0.64.0). A call opened with audioTags: true
lets the voice ACT OUT these cues, and the provider's transcript still
contains them; connectCall strips them before onTranscript and before
the transcript reaches memory. Exported for integrations that open the
ElevenLabs session themselves from startCall credentials.
Only ASCII-word cues are removed (the cue vocabulary is English): a line's
own brackets — 【重点】, [1] — are left alone. Padded ([ laughs ]),
full-width (［laughs］) and trailing-ellipsis ([sighs…]) spellings of a
cue count.
Remove Eleven v4 performed cues ("[laughs]", "[whispers]", "[short pause]") from a call transcript line (0.64.0). A call opened with
audioTags: truelets the voice ACT OUT these cues, and the provider's transcript still contains them;connectCallstrips them beforeonTranscriptand before the transcript reaches memory. Exported for integrations that open the ElevenLabs session themselves fromstartCallcredentials.Only ASCII-word cues are removed (the cue vocabulary is English): a line's own brackets —
【重点】,[1]— are left alone. Padded ([ laughs ]), full-width (［laughs］) and trailing-ellipsis ([sighs…]) spellings of a cue count.