The default Dataset contains one item for each unique input video ID. All fields below are present in both successful and error rows unless noted as nullable.
| Field | Type | Always present | Description |
|---|---|---|---|
videoId |
String | Yes | Normalized YouTube video ID. |
videoUrl |
URI string | Yes | Canonical public watch URL. |
status |
success or error |
Yes | Whether a complete transcript was delivered. |
title |
String or null | Yes | Public title when available. null means the title could not be recovered. |
languageRequested |
String | Yes | Validated language code sent for this result. |
segmentCount |
Integer | Yes | Number of items in segments; zero for error rows. |
segments |
Array | Yes | Timestamped transcript segments. Empty for error rows. |
segments[].start |
Number | For segment rows | Start time in seconds. |
segments[].duration |
Number | For segment rows | Segment duration in seconds. |
segments[].text |
String | For segment rows | Caption text. |
errorCode |
String or null | Yes | Stable classification such as NO_TRANSCRIPT or LANGUAGE_UNAVAILABLE. |
errorMessage |
String or null | Yes | User-facing explanation for an error. |
Nullable fields are explicit: title can be unavailable, while errorCode and errorMessage are null on success. A successful item is billable only when segments is non-empty and complete. Error items remain useful for auditing failed inputs but are not billable transcript events.