The Free Google Tool That Reads All Your Docs (NotebookLM)

Ask a general AI chatbot a question about your own business and it will still answer, even when it has never seen a single one of your documents. It's just guessing from patterns, confidently. NotebookLM works the opposite way. You upload your own files first: contracts, manuals, meeting notes, reports, whatever you want it to know, and it will only ever answer from what's actually in there. If the answer isn't in your documents, it says so instead of making something up.

What NotebookLM actually is

NotebookLM is a free tool from Google. You create a "notebook," upload a set of your own documents as sources, and then ask it questions the way you'd ask a colleague who has actually read everything you gave them. Every answer comes back with citations pointing to the exact source and passage it pulled from, so you can check its work in one click instead of taking it on faith.

That grounding is the whole point. A general chatbot answers from its training data, which was never your business. NotebookLM answers from your business, because that's the only thing it's allowed to look at.

Worth knowing: Google renamed the product to Gemini Notebook in July 2026. Same tool, same core idea, new name folded into the wider Gemini family. Most people still search for it as NotebookLM, so that's the name this post uses, but don't be thrown if the interface you land on says Gemini Notebook instead.

What you can actually feed it

You're not limited to plain text. NotebookLM accepts PDFs, Google Docs, Google Slides, Word documents, websites, YouTube videos, audio files, and text you paste in directly. Drop in a whole folder of onboarding manuals, a year of board meeting notes, or a stack of client contracts, and it treats all of it as one connected set of sources you can question together.

That's a genuinely different way of working than searching one file at a time. Instead of hunting through twelve PDFs for the one clause you half-remember, you ask the notebook directly and it tells you which document it's in.

The feature everyone talks about: Audio Overviews

The feature that made NotebookLM famous is Audio Overviews: it can turn your uploaded sources into a spoken conversation between two AI hosts discussing what's actually in your documents. There are a few formats to pick from, not just the one style: a longer two-host "deep dive" discussion, a short brief where a single voice delivers the key points in under two minutes, a critique format that gives constructive feedback on something like a draft or a proposal, and a debate format where two hosts argue different sides of the material.

For a business owner, the practical use is simple: turn a dense document into something you can listen to on a commute or a walk, instead of setting aside dedicated reading time you don't have.

Where this is genuinely useful for a small business

A few fictional examples to make the shape concrete, not real clients:

• A physiotherapy clinic uploads its entire staff handbook and standard operating procedures, then lets new hires ask the notebook questions directly during onboarding instead of interrupting a manager for the tenth time that week.
• A boutique accounting firm uploads a client's full set of financial statements and lets a partner ask targeted questions before a review call, with every answer citing the exact statement and line it came from.
• A coaching business uploads transcripts from several client calls and generates an Audio Overview to review common themes on a drive between sessions, instead of re-reading pages of notes.

The pattern across all three: a pile of documents nobody has time to fully read, turned into something you can question or listen to instead.

What it isn't

NotebookLM only knows what you put in front of it. It doesn't watch your inbox, it doesn't pull fresh data from your calendar, and it won't know about anything that happened after you uploaded your sources. For that kind of always-current, connected AI, that's a different job, the one Claude's Gmail and Calendar connectors are built for. NotebookLM and connectors solve different problems: one is a deep, grounded expert on a fixed set of documents, the other is an assistant that can see what's happening in your business right now.

It's also worth pausing before uploading anything with client-identifiable information in it. If a document contains names, account numbers, or anything sensitive, treat it the same way you would before handing it to any AI tool: strip or anonymise what doesn't need to be there first. This is exactly the problem the redaction workflow solves if you're working with client files regularly.

And unlike Claude's memory feature, which quietly builds up context about you across ordinary conversations over time, NotebookLM's "memory" is only ever the documents you deliberately chose to upload into that specific notebook. Nothing carries over unless you put it there yourself.

FAQ

Is NotebookLM actually free?

Yes. There's no credit card required and no trial clock running out, you sign in with a Google account and start uploading sources. Google also offers paid tiers with higher limits for heavier use, but the core tool, including Audio Overviews, works on the free plan.

Is NotebookLM the same as Gemini Notebook?

Yes. Google renamed NotebookLM to Gemini Notebook in July 2026 as part of pulling the product closer into its Gemini lineup. It's the same tool underneath: existing notebooks, sources and links kept working through the rename, nothing needed to be rebuilt.

What file types can I upload to NotebookLM?

PDFs, Google Docs, Google Slides, Word documents, plain text, websites, YouTube videos, audio files, and text you paste in directly. You can mix different types of sources in the same notebook and question them all together.

How is this different from just asking Claude or ChatGPT about my documents?

A general chatbot answers from everything it was trained on, so it will still guess even if it has never seen your files. NotebookLM only answers from the sources you've actually uploaded into that notebook, and every answer comes back with a citation pointing to exactly where it found it. If you want the same "grounded, not guessing" behaviour inside your everyday AI chat instead of a separate notebook, Claude's connectors work toward a similar goal by linking straight into your live Gmail and Calendar data.

If you've got a folder of manuals, contracts or meeting notes nobody has properly read in months, that's exactly what NotebookLM is for. Book a strategy call and we'll find the first set of documents worth turning into a notebook.