Class VoiceRecordingRuntime

constructor

• new VoiceRecordingRuntime(init?: VoiceRecordingInit): VoiceRecordingRuntime

  Parameters

  • init: VoiceRecordingInit = {}

  Returns VoiceRecordingRuntime

Readonlysegments

Readonlytimeline

duration

• get duration(): number

  Total duration in seconds (max segment endTime).

  Returns number

fullWav

• get fullWav(): Uint8Array

  Full mixed/concatenated conversation audio as WAV bytes.

  Returns Uint8Array

save

• save(path: string, format?: string): string

  Save the full conversation to a file.

  Format is inferred from the path suffix or overridden by format. .wav is written natively; non-WAV formats transcode via the bundled ffmpeg subprocess (see resolveFfmpegPath).

  path is resolved before writing; format is validated against an allowlist so callers can't pass arbitrary ffmpeg muxer names.

  Parameters

  • path: string
  • Optionalformat: string

  Returns string

saveSegments

• saveSegments(dir: string, options?: { manifest?: boolean }): string

  Write each segment as its own WAV file plus the full mixed conversation, optionally with a JSON manifest pairing files to transcripts/timestamps.

  Layout:

  <dir>/
      segments/
          00-user-0000ms.wav
          01-agent-0312ms.wav
          ...
      full.wav
      manifest.json   # iff manifest=true
  

  Segment file names: zero-padded index, role, start_time in milliseconds. Existing files inside dir are NOT cleared — caller decides retention.

  Parameters

  • dir: string
  • options: { manifest?: boolean } = {}

  Returns string