Skip to content

assist transcript

See Meeting transcripts for processing a folder of transcripts with /transcripts.

  • assist transcript configure asks for the pick-up, transcripts and summary folders.
  • assist transcript clean <path> prints a .vtt file as a readable markdown transcript, one paragraph per speaker turn. --timestamps starts each turn with its time, and --format vtt prints a tidied .vtt instead.
  • assist transcript merge <path...> --out <file> combines several .vtt files into one, on a single timeline.
    • --select <file> keeps only the passages a JSON file names, as { "keep": [{ "file", "from", "to" }], "removed": ["reason", ...] }, with times as --timestamps prints them. - reads it from stdin.
    • --no-provenance leaves out the notes naming the source files, where each passage came from and how many were removed. Use it when the merged transcript goes to people who shouldn’t see those.
    • --widen-audience removes the asides that only make sense to people who were in the call.
  • assist transcript list internal prints the .vtt files waiting in the pick-up folder. Called by /transcripts.
  • assist transcript move <file> --date <YYYY-MM-DD> --client <name> internal writes a pick-up file as a dated markdown transcript, archives the original and prints where its summary goes. Called by /transcripts.