audio-metrics

The audio-metrics transform walks all upstream byte blobs, attempts WAV/AIFF (stdlib) or mutagen parsing, and emits one JSON document listing every file with metrics or a parse error.

When to use it

Use after convert-to-pcm (or directly on raw uploads) when downstream nodes need a stable files array keyed by MD5 (for example generate-transcript and voice-analysis).

Limitations

• Mixed DAG inputs may include non-audio blobs; those appear with isValidAudio: false and parseError rather than failing the node.
• Metrics are descriptive (duration, channels, sample rate, container), not a dashboard or analytics product by themselves.
• Does not transcribe or classify content; pair with generate-transcript for text.

Configuration (summary)

Output shape

{
  "audioMetricsVersion": 1,
  "generatedAt": "2026-01-01T00:00:00+00:00",
  "files": [
    {
      "md5": "...",
      "sizeBytes": 12345,
      "sourceNodeName": "convert-to-pcm",
      "index": 1,
      "isValidAudio": true,
      "container": "wav",
      "durationSeconds": 12.5,
      "channels": 1,
      "sampleRateHz": 48000,
      "bitDepth": 16
    }
  ]
}

Invalid items keep md5 and parseError with isValidAudio: false.

Related

• Audio Utils
• convert-to-pcm
• generate-transcript