Speakers in a YouTube transcript
WithTranscript marks speakers, but never by name. Some transcripts label each paragraph Speaker A, Speaker B and so on; transcripts built from a video's YouTube captions only mark where the speaker changes, the way the captions do. To know who's who, use the video. Part of getting a YouTube transcript.
Labelled speakers
In transcripts with labels, each paragraph carries a letter for its speaker, and the copied text keeps it:
**[0:01] Speaker A:** Welcome back.
**[0:09] Speaker B:** Thanks for having me.The letters are consistent within a video, so Speaker A is the same voice throughout, but they're not linked to names. In a podcast, the host usually introduces the guest in the first minute; that's the easiest way to match letters to people.
Transcripts built from captions
New transcripts are built from the video's captions. Captions often mark a change of speaker (on YouTube you'll sometimes see >>) but don't say who it is. WithTranscript keeps those turn marks: in the copied text a paragraph that starts a new turn begins with a dash.
**[4:10]** — So what changed?
**[4:12]** — Honestly, the tooling.Quoting someone accurately
Before you attribute a line, click it in the transcript and listen. It takes a few seconds and catches the cases where a turn mark is off by a sentence. Quoting a podcast covers this in more detail.
Example from the library
Ben & Marc: Why Everything Is About to Get 10x Biggera16z · 54 minA two-person conversation in 7 chapters.