> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lavendly.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Get audio summary

> Derived per-clip audio model. Single source of truth = (nodes, edges).

Returns the audio state of every `videoGenerator` clip in the workflow.
The `audio_tracks` array is **derived** from the workflow's nodes and edges
on every read, agents that attach tracks via [POST tracks](/api-reference/audio/attach-track)
see them here immediately without waiting for the canvas to run propagation.

<ParamField path="id" type="string" required>Workflow id.</ParamField>

### Response

```json theme={null}
{
  "workflow_id": "wf_xyz",
  "schema_version": 1,
  "clips": [
    {
      "clip_id": "vid1",
      "prompt": "...",
      "duration": 5,
      "native_audio": { "mode": "mix", "volume": 1 },
      "audio_tracks": [
        {
          "track_id":       "tr_a1b2c3d4e5f6",
          "source_node_id": "vo_xyz",
          "target_clip_id": "vid1",
          "kind":           "voiceover",
          "url":            "https://.../voice.mp3",
          "startTime":      0,
          "duration":       5,
          "volume":         1,
          "muted":          false,
          "ducking":        false,
          "subtitleStyle":  "tiktok",
          "transcribedWords": [{ "word": "Hello", "start": 0.1, "end": 0.4 }]
        }
      ]
    }
  ]
}
```

### native\_audio modes

| Mode     | Meaning                                                      |
| -------- | ------------------------------------------------------------ |
| `off`    | Strip the clip's baked audio entirely.                       |
| `native` | Use only the video model's baked dialogue and sound effects. |
| `mix`    | Blend the model's native audio with the attached tracks.     |