To change a NotebookLM Audio Overview's duration, open its generation controls and choose Shorter, Default, or Longer before generating. Google currently documents those length choices for English Audio Overviews. They are presets, not a field where you enter an exact number of minutes.
Start with NotebookLM's native control. You can also change the format, select fewer sources, or add instructions that narrow the topic. Use Jellypod only when you need to edit the script line by line, choose the voices, or keep a repeatable runtime across a hosted series. Google's current Audio Overview help is the source of truth for the native length options and notes that they are English-only.
Open Audio Overview in the Studio panel, then set the format, language, and length before generation.
Google renamed NotebookLM to Gemini Notebook on July 16, 2026. It is the same standalone product, and the Audio Overview steps below apply to Gemini Notebook.
Where are the Shorter, Default, and Longer controls?
The length setting appears in the Audio Overview generation panel. Open or create a notebook, add the sources, select Audio Overview in Studio, and open its customization controls. Choose one of the three length presets, then generate the audio.
Google's Audio Overview instructions describe the choices this way:
| Setting | What it tells NotebookLM | Use it when |
|---|---|---|
| Shorter | Generate a shorter version of the chosen Audio Overview format | You want a quick pass but still want that format's structure |
| Default | Use NotebookLM's standard length for the selected format | You have no fixed runtime requirement |
| Longer | Generate a longer treatment of the selected sources | The default pass leaves useful context out |
Google does not publish a guaranteed minute range for these three presets. Treat them as relative instructions. The source set, format, topic, and generation can all affect the finished runtime, so listen to the result rather than planning a schedule around an assumed number.
Choose The Brief for a documented single-speaker recap under two minutes. Use the English length presets for relative changes and an editable script for production constraints.
Do the length presets work in every language?
No. Google's help marks Shorter, Default, and Longer as English-only, even though NotebookLM can generate Audio Overviews in more than 80 output languages. Do not assume the same control appears when the output language is set to German, Spanish, or another supported language.
For a non-English overview, choose the best format and use a narrow focus instruction. Keep the selected source set tight, then review the generated audio. If a fixed runtime is part of an assignment, campaign, or training module, an editable script gives you more control than repeated generation.
Is The Brief the same as choosing Shorter?
No. The Brief is a format, while Shorter is a length preference. Google's documentation says The Brief uses one speaker to deliver the document's key takeaways in under two minutes. Deep Dive, The Critique, and The Debate use different structures and purposes.
Choose The Brief when a sub-two-minute, single-speaker recap fits the job. Choose Shorter when you want a shorter version of another format, such as a two-host Deep Dive. That difference matters because switching formats can change the voice count and editorial shape, not only the clock.
| Your goal | Better native choice |
|---|---|
| A single-speaker recap under two minutes | The Brief format |
| A shorter two-host explanation | Deep Dive with Shorter selected |
| More context around connected source themes | Deep Dive with Longer selected |
| Concise feedback on a document | The Critique with a narrow focus prompt |
These are workflow recommendations based on the format definitions in Google's official help page, not guaranteed runtime benchmarks.
Can a prompt set an exact NotebookLM runtime?
NotebookLM lets you enter instructions that focus on specific topics or adjust the expertise level. Google documents the three length presets separately and does not document an exact-minute input in the generation controls.
Use the prompt to control scope rather than make a timing promise. Name the listener and the content that must survive a shorter cut:
Create a concise orientation for new faculty. Cover the revised grading deadline and exception process. Skip the history section and define the two policy acronyms.
That prompt gives NotebookLM a clearer selection problem. It does not guarantee a specific duration. If the first version is long, remove sources that are outside the task, narrow the focus, choose Shorter, and generate another overview. If a sentence must stay or go, move to an editable script instead of relying on another generated take.
Does playback speed change the duration of the audio file?
No. Playback speed changes how quickly you hear the Audio Overview in NotebookLM. It does not rewrite the content or create a shorter source file. Google lists Change playback speed as a player control, separate from the length preferences used during generation.
Use playback speed when the listener wants to finish sooner. Use Shorter, Default, or Longer when the content itself needs a different amount of detail. If you download the original overview after listening at a faster speed, the generation itself has not been shortened.
How can source selection make an Audio Overview shorter?
NotebookLM generates artifacts from the sources selected in the notebook. Google's source guide recommends selecting a specific set of sources when you want the model to use only part of a notebook.
For a shorter overview, remove material that does not answer the listener's question. If a notebook contains a syllabus, twelve readings, meeting notes, and a policy appendix, do not select all of them for a three-minute orientation. Select the policy and the approved summary. Then tell NotebookLM which decision or change the listener needs to understand.
This is an editorial choice, not a mechanical one. A short source can still need explanation, and a long document can contain one narrow section that produces a concise overview. Source relevance is a better filter than page count.
How do you review a shorter or longer generation?
Check meaning before minutes. Google warns that Audio Overviews and their voices may contain inaccuracies or audio glitches. A shorter generation can also omit context you expected, while a longer one can spend time on a secondary theme.
Review the result against a small checklist:
- Did it answer the listener's main question?
- Did it preserve names, dates, conditions, and exceptions that affect the meaning?
- Did it include the topic named in the focus instruction?
- Did any audio glitch or speaker switch make a statement unclear?
- Is the finished runtime acceptable for where it will be used?
Do not remove a qualification from medical, legal, financial, or policy material just to hit a shorter target. Revise the scope or split the material into separate overviews instead.
When do you need an editable script instead of another preset?
The native presets are right for relative changes: make this generation shorter, leave it at the default, or let it run longer. An editable script is better when the runtime sits inside a production constraint, such as a course module, an internal training slot, or a recurring show with a consistent format.
Jellypod adds the production controls that come after a source-based overview.
In Jellypod, you can revise exact lines before publishing, remove a section without asking for a new interpretation, assign approved voices, and preview the finished episode. The same show can keep its hosts and structure across later episodes. Hosting and RSS distribution then carry that repeatable production into a series listeners can follow.
That is a different job from making a quick Audio Overview. Use NotebookLM for the native summary. Use Jellypod when the script, voices, and publishing system need to stay under your control.
Can you change the length of an Audio Overview you already generated?
The length choice belongs to generation, not playback. For different content length, return to the Audio Overview controls, change the preference or focus, and generate another version. Keep the first version until you have checked that the new one preserves the details you need.
If the existing audio is already close, importing the NotebookLM file into Jellypod can turn it into an editable script. That route makes sense when you want to cut or expand exact passages, change voices, or publish the result through a hosted feed.
The practical next step
Choose Shorter, Default, or Longer in NotebookLM and listen to the full result. If relative length is enough, stop there. If exact wording, voice choice, hosting, or a repeatable episode structure matters, move the approved source material into an editable production workflow.