How to get a podcast into Gemini Notebook
Gemini Notebook (formerly NotebookLM) takes text, not audio feeds, so a podcast has to be transcribed first. The part people get wrong is what happens next: uploading one file per episode burns the source limit fast, and loses the episode titles and dates Gemini Notebook needs to cite anything usefully. Transcribe each episode, keep the title, date, and link as a header inside the text, then merge the episodes into a small number of large documents rather than one file each.
Step by step
-
1
Get a transcript for each episode
Some shows publish transcripts — check the show notes or the RSS feed first, because that's free and already accurate. For everything else you need speech-to-text. Whisper and its variants are the usual choice and handle podcast audio well.
-
2
Keep the metadata inside the text
Put the episode title, date, and source URL as a short header at the top of each transcript. If that lives only in the filename, Gemini Notebook's answers come back unattributable — which defeats the reason for using it over a plain chat box.
-
3
Merge episodes into a few large documents
Gemini Notebook caps you on number of sources, so 200 episodes as 200 sources is a non-starter. Concatenate them into a handful of documents packed toward the word limit, with a clear separator and header between episodes.
-
4
Split on topic boundaries, not word counts
If you have to split a long episode, cut where the subject changes rather than at an arbitrary word count. Cutting mid-explanation is the most common cause of answers that are confidently half-right.
-
5
Upload and check a citation
Ask a question you already know the answer to and confirm the citation points at the right episode. If it can't name the episode, the metadata didn't survive step 2.
Doing it by hand
The whole process without any particular tool, if you'd rather build it yourself:
- Download the episode audio you have the rights to use.
- Run it through Whisper (locally, or a hosted API) to get text.
- Paste a title/date/URL header at the top of each transcript.
- Concatenate the transcripts into a few large .md or .txt files.
- Upload those files to Gemini Notebook as sources.
Common questions
Can Gemini Notebook read an RSS feed directly?
No. It accepts documents, pasted text, Google Docs, and some web URLs, but it doesn't subscribe to a podcast feed or transcribe audio for you. The transcription step is yours.
How many episodes fit in one notebook?
It depends on episode length and the current source and word limits, but merging is what decides it. As individual files you run out of sources long before you run out of words; merged into large documents, a few hundred hours of audio fits comfortably.
Do I need speaker labels?
For a multi-guest show, yes — a wall of undifferentiated text makes it much harder to attribute a claim to the right person, and quotes come back mixed between speakers. For a solo show it adds little.
The shortcut
SourceWeaver does every step above in one pass — reusing existing captions and transcripts, transcribing the rest on a GPU with optional speaker labels, and exporting in the format your destination tool actually wants. You can see real output before signing up.
Process content you own or created, content you're licensed or have permission to use, or openly-licensed and public-domain material. Whether any particular use is permitted depends on copyright law and the rights you hold — it's worth being sure before you start.