LM Subs blog
Practical, tested guides about captions, subtitles, and transcript export for Google NotebookLM Video Overviews and Audio Overview podcasts. Written and maintained by Sarang (@sarang-cmd), the creator of LM Subs.
9 guides, updated for LM Subs v0.2.8.
Captions basics2 min read Does Google NotebookLM have captions or subtitles? Straight answer: NotebookLM Video Overviews and Audio Overviews ship without closed captions. Here is what the built-in player does and does not offer, and how to add real timed subtitles. Read the guide
Setup
Add subtitles to Video Overviews
A complete walkthrough: install the extension, add one free speech-to-text key, play a Video Overview, and export timed captions as .SRT, .VTT, or Markdown notes.
API keys
Free speech-to-text API keys
Twelve bring-your-own-key speech-to-text providers compared by free tier, timestamp quality, and setup friction, with the exact page to visit for each key.
Transcript export
Export NotebookLM transcripts
Which subtitle format to use for video editors, web players, study notes, spreadsheets, and Anki decks, plus how timing offsets and speaker labels are preserved on export.
Audio Overviews
Audio Overview podcast captions
Audio Overviews hide their audio element inside a detached player. Here is how captions attach to the podcast format, including speaker labels and the caption dock.
Transcript export
SRT, VTT, or Markdown?
Every export format LM Subs writes, what it is actually for, and which one to pick for video editors, slide decks, study notes, or spreadsheets.
Local setup
Run Whisper locally
Point LM Subs at a local Whisper server and caption NotebookLM Video Overviews and Audio Overviews entirely on your own machine, with no account and no audio leaving the computer.
Costs
What captions actually cost
A clear breakdown of what LM Subs sends to a speech to text provider, which free tiers cover it, and the point at which paying makes sense.
Accessibility
Captions and WCAG
What WCAG 2.2 requires for captions and transcripts, where NotebookLM falls short, and how to close the gap for deaf, hard of hearing, and neurodivergent users.