Studio Feature · Audio1 free

Audio Overview — The AI Podcast Engine

Audio Overview is the feature that made NotebookLM famous. It takes your uploaded documents — research papers, meeting transcripts, textbook chapters, business reports — and transforms them into a podcast-style conversation between two AI hosts who explain, debate, and contextualize your material. The hosts use analogies, ask each other questions, and make complex information feel like something you'd hear on a commute.

What it does and why it matters

Audio Overview generates an 8–20 minute two-host podcast from your source materials. Unlike text-to-speech, the AI hosts actually discuss the content — they summarize key arguments, highlight surprising findings, debate implications, and use everyday analogies to make dense material accessible. The feature launched in September 2024 and quickly became NotebookLM's most viral feature, with educators, researchers, and content creators sharing their AI-generated podcasts widely.

The hosts don't just read your documents aloud. They synthesize across all sources in the notebook, identify the most interesting threads, and construct a narrative arc. This means a notebook with five research papers produces a podcast that connects themes across all five papers — something that would take a human host hours of preparation.

Since December 2024, you can join the conversation by tapping the "Join" button during playback, letting you ask follow-up questions and steer the discussion in real time. Audio is now available in over 80 languages, and you can customize tone (formal, casual, humorous) and focus areas through the prompt field.

When to use Audio Overview

Audio Overview excels in situations where you need to absorb information without reading. Commute learning is the most common use case — upload a 40-page report before you leave the office, and listen to a 15-minute summary on the drive home. Pre-meeting preparation is another strong case: generate a podcast from the meeting's background documents and listen while you get ready.

It's also remarkably effective for testing your own understanding. Upload a document you've written — a thesis chapter, a report draft, a proposal — and listen to two AI hosts explain it back to you. You'll immediately hear where your argument is clear and where it's muddled. This "listen to your own work" technique has become popular among academic writers.

For team alignment, sharing an Audio Overview is faster than asking everyone to read a long document. A product manager can upload a PRD and competitive analysis, generate a podcast, and share the link — the team gets context in 10 minutes instead of an hour of reading.

Limitations to know

Audio Overviews cannot be edited after generation — if you don't like the result, you regenerate. Each generation may produce slightly different content, which can be a feature (different perspectives) or a limitation (no consistency). Very long documents may be truncated in the context window, so the hosts sometimes gloss over material from later sections. For best results, keep notebooks focused on a single topic and limit source count to 10–15 documents.

Step-by-step workflow

6 steps
01

Upload your source materials

Add 1–50 documents to a focused notebook. For best podcast quality, use 3–10 sources on a single topic. PDFs, Google Docs, web URLs, and YouTube transcripts all work. Name your sources clearly — the hosts may reference them.

02

Open the Studio panel

Click the Studio panel on the right side of the NotebookLM interface. You'll see all available output formats. The Audio Overview tile shows a speaker icon.

03

Select sources to include

You can generate audio from all sources or select specific ones. For a focused podcast, select just the 3–5 most relevant sources rather than everything in the notebook.

04

Customize your generation prompt

Click the pencil/edit icon on the Audio Overview tile. In the prompt field, specify: what the hosts should focus on, the target audience level (beginner, expert), the desired tone (casual, academic, humorous), and any specific questions you want answered.

05

Generate and wait

Click Generate. Audio Overviews typically take 2–5 minutes depending on source volume. Longer documents with more complex material take longer. You can work on other tasks while it generates.

06

Listen, join, and iterate

Play the podcast. Use the 'Join' button to enter the conversation and ask follow-up questions. If the output misses key points or emphasizes the wrong things, adjust your prompt and regenerate. Download the audio for offline listening.

Prompts

1 free

Core Generation

1 prompt
Generate an Audio Overview focused specifically on the methodology and research design described in these sources. Have the hosts walk through the study design step by step, explain why each methodological choice was made, discuss the strengths and potential weaknesses, and compare the approach to alternatives. Target audience: graduate students learning research methods.

Advanced Use Cases

1 prompt
Unlock All Prompts

Get the complete prompt library for this category.

Every prompt in this guide plus all prompts across the full category — advanced workflows, specialized use cases, and production-grade templates.

Category Bundle — one-time access

Unlock Category Prompts — $19.99

ONE-TIME · 30-DAY GUARANTEE · INSTANT ACCESS

Related Guides
Create a Podcast Customize Audio Audio & Interactive Mode Audio Masterclass Custom Podcast Styles
← Back to All Guides