From an earlier video

This is material from an earlier video — the models, prices and limits of AI services have changed since. The idea and the approach still hold.

TL;DR

  • Google’s NotebookLM turns your Obsidian notes into interactive podcasts and helps you find gaps in your knowledge.

  • How to use it: export folders via the Advanced Merger plugin, upload them to NotebookLM, get audio podcasts and an analysis of your thinking.

  • Main plus: it works only with your documents, without hallucinations.

  • Main minus: there’s no auto-sync — every note update requires a manual re-export.

  • Who it suits: for deep analysis of large sets of notes, exam prep, creating educational content.


What NotebookLM can give an Obsidian user

NotebookLM is an AI tool from Google that works exclusively with your documents. Unlike ChatGPT, it doesn’t make up facts — it only analyses what you’ve uploaded.

Key capabilities for Obsidian:

  • Audio podcasts from notes — turns dry summaries into a dialogue between two AI hosts
  • Knowledge-gap analysis — shows which topics you avoid or misunderstand
  • Checking insights — compares your notes with the original sources
  • Mind maps — visualises the connections between concepts from different notes

How to connect it: exporting via Advanced Merger

NotebookLM works with whole documents, while Obsidian creates lots of short notes.

The solution — combine related notes into files.

Installing Advanced Merger

  1. Settings → Community plugins → turn off Safe mode
  2. Browse → find “Advanced Merger” → Install → Enable
  3. Restart Obsidian

Exporting notes

  1. Organise by topic: create folders for related notes (e.g. “AI research”, “Productivity books”)
  2. Right-click on the folder → “Merge folder”
  3. Export to PDF — use Better Export PDF for large volumes
  4. Check the size — files up to 200MB work more reliably

Important: internal links [[note]] are lost on export. So it’s better to export self-contained topic blocks.


Four practical ways to use it

1. Podcasts from study materials

Upload your lecture or book summaries to NotebookLM → click “Generate Audio Overview”. You get a 10–15 minute podcast dialogue where the AI discusses the key ideas.

Case: a student uploaded their quantum-physics notes and got a podcast they listened to on the way to university to reinforce the material.

2. Finding blind spots in your knowledge

Upload all your notes on the topic you’re studying and ask: “Which areas do I constantly avoid or under-develop?”

Example: a developer discovered they avoid security topics, even though it’s critical for their projects.

3. Checking the accuracy of your summaries

Upload the original sources + your notes. Ask: “Compare my summaries with the original sources. What did I distort or miss?”

Handy for: students before exams, researchers writing papers.

4. Creating study materials

Use the Study Guide to automatically create structured study aids from chaotic notes.

Downsides and limitations (honestly)

No synchronisation — the main problem

A critical flaw: changes in Obsidian don’t reach NotebookLM automatically. Every note update requires:

  1. Re-exporting the folders via Advanced Merger
  2. Re-uploading to NotebookLM
  3. Losing the chat history with the previous version

In practice: if you actively develop your notes, you’ll have to update NotebookLM weekly.

Loss of the graph structure

Obsidian’s unique strength — the links between notes — is completely lost. NotebookLM sees only the text, it doesn’t understand the context of [[note]] links.

Technical limitations

  • A maximum of 50 sources per project (too few for large vaults)
  • File size up to 200MB on upload
  • Podcasts only in English (Russian text is supported)
  • Requires the internet — doesn’t work locally

The workflow becomes fragmented

The “Obsidian → export → NotebookLM → analysis → back to Obsidian” process breaks the natural flow of working with notes.

Who it suits, who it doesn’t

✅ Worth a try if you:

  • Are a student or researcher with large amounts of study material
  • Are preparing for exams and want to find gaps in your knowledge
  • Create educational content based on notes
  • Work with 100+ notes on a specific topic
  • Need podcasts for mobile learning

❌ Not for you if you:

  • Update your notes often and need up-to-date information
  • Work with confidential data (everything is stored in Google’s cloud)
  • Get the main value of your vault from the links between notes
  • Need Russian-language audio content
  • Prefer local solutions

Final recommendation

NotebookLM is a powerful tool for periodic deep analysis of large blocks of notes, but not for daily work in Obsidian.

The optimal strategy: use it once a month to analyse accumulated knowledge on a topic, create podcasts from summaries or prepare for important events.

Start with an experiment: pick one topic folder (50–100 notes), export it via Advanced Merger and test the features. It’s free and will show whether the tool fits your workflow.


Next article: Claude MCP + Obsidian: direct integration without exporting files — learn how to connect Claude directly to your vault and work with notes without manual export.