The file's name or path.
The file's bytes or its text, whole where it can be: a Word file, SubRip and WebVTT are named from their first 4 KB, but Fireflies's JSON can hold its sentences, and Otter's export its second turn or its closing line, past any head of the file. A head's last line with no line feed after it may stop short, so it is read as words, never as a header.
The format of a transcript file, by its name's extension and its bytes, which must be the format the name claims;
nullwhen it is none of them. A.srtholds a timing line; a.vttopens withWEBVTT; a.jsonholds Fireflies's keys; a.docxis a ZIP archive; a.txtor.mdis text,otterwhen it is laid out as Otter's export andlinesotherwise. The bytes are read as readTranscript reads them. A PDF is no transcript, whatever its name.