Interface TranscriptCue

A single timed transcript cue — the unit produced by ASR and by MJ's realtime session capture (which records speaker + timings per turn).

interface TranscriptCue {
    EndMs: number;
    Speaker?: string;
    StartMs: number;
    Text: string;
}

Properties

EndMs: number

Cue end, milliseconds from the beginning of the asset.

Speaker?: string

Optional speaker label/id — a speaker change is a strong boundary signal.

StartMs: number

Cue start, milliseconds from the beginning of the asset.

Text: string

Spoken text for this cue.