Indexing closed caption files: SAMI, TTML or WebVVT

These files have timestamps mingled in between the words. Is it possible to index them to return the nearest timestamp to a given word/phrase?