Transcript vs Subtitles vs Captions on YouTube: What's the Difference?
· 4 min read
On YouTube these words point at more or less the same underlying data, which is why using them loosely rarely causes problems. But they are not identical, and the differences matter when you are exporting, translating, or working on accessibility.
Captions
Text synchronised to the audio, intended primarily for viewers who cannot hear it. Closed captions conventionally include non-speech information — speaker labels, and cues like [door slams] or [laughter] — because they are meant to convey the full audio experience, not just the words.
On YouTube, captions are either uploaded by the creator or generated automatically by speech recognition. Auto-captions are speech-only: they will not tell you a door slammed.
Subtitles
Strictly, subtitles assume the viewer can hear the audio but may not understand the language — so they carry dialogue and little else. In practice, and in YouTube's own interface, 'subtitles' and 'captions' are used almost interchangeably, and the distinction only really holds in professional broadcast and localisation work.
Transcript
The same text, laid out for reading rather than for display over video. A transcript is continuous and scannable; captions are chunked into short timed lines sized to fit on screen.
YouTube's transcript panel is generated from the caption track — which is why a video with no captions also has no transcript. They are the same source data presented two ways.
SRT and VTT files
These are file formats for caption data, not a different kind of content. An SRT file is a plain text file listing numbered entries, each with a start and end timecode and its line of text. WebVTT is a similar format built for the web with some extra styling capability.
You need one of these formats when you are uploading captions to a platform or importing them into video editing software. For reading, quoting, or feeding text to an AI tool, a plain transcript is more useful — the timecode scaffolding just gets in the way.
When the distinction matters
- Accessibility compliance: you likely need true closed captions with non-speech cues, not just an auto-generated speech transcript.
- Re-uploading or localising: you need a timed format like SRT or VTT.
- Reading, research, quoting, or AI analysis: you want a plain transcript.
- Editing workflows: your software will specify the format it expects.
Summarize and chat with any YouTube video
Free transcripts on the web, plus AI summaries and chat right on the YouTube page with the free Chrome extension.
Add to Chrome — Free