Transcript Exact Match

Performs an exact text match search on video transcripts using a natural language text query (e.g., 'thank you for your time' or 'my name is John') within a specified dataset (dataset_id). This searches for exact occurrences of the provided text within video transcript data, not semantic or approximate matching. Returns videos grouped by their exact transcript matches, where each match includes the composite slice (shot or scene) containing the match, the matched transcript text, timestamps, frame numbers, and a coverage score indicating how much of the composite's transcript is covered by the query (1.0 means the query matches the entire transcript).