Blue title graphic with the German words KI fürs Studium, meaning AI for studying

Studying

AI tools for studying: NotebookLM or Astra AI?

By Daniel Ohiggins · Published: November 17, 2026 · 8 min read

You have a set of lecture notes and need a revision sheet. Should you pay for an AI subscription, use NotebookLM or consider Astra AI? The question “Astra AI or NotebookLM?” from 15 October is a useful starting point. What matters, though, is what you want to do with your material. Fluent prose does not tell you whether you understand the subject.

What the study assistants share and where they differ

Both products are candidates for study support. That does not mean they have identical features. NotebookLM works with supplied sources: you provide material and ask it to produce summaries, for example. Audio Overviews present content for listening. The research baseline calls the product Gemini Notebook; we use the familiar name NotebookLM here.

The same research describes Astra AI as an app-based study assistant with a subscription model. Specific feature limits and prices are missing. A comparison table suggesting equivalent features would therefore mislead readers. We have not conducted a first-hand performance test and do not name either product a test winner.

Start with the source question: should the answer reflect a particular set of lecture notes, or do you need an open-ended explanation? Even a source-grounded answer can misrepresent something. Keep the original open beside the result. A reference helps only if the cited passage actually supports the statement.

Prices and limits: mid-November 2026, please check

Editorial date: mid-November 2026. The figures below come from the project plan verified on 15 October 2026. They are not a fresh pricing check on the publication date. Check prices and usage limits with the provider before paying or making your exam preparation depend on them.

NotebookLM / Gemini NotebookPrice in the research baselineNotebooksSources per notebook
Free€010050
Google AI Plus€4.99 per month200100
Google AI Pro€21.99 per month500300

The same baseline lists 50 chats per day for free access. It also lists audio summaries in more than 80 languages, including German. Additional chat or audio limits for the paid plans are not documented here. A price does not imply unlimited use.

Specific Astra AI pricing remains unverified. Before subscribing, check the price of your actual offer, billing period and cancellation terms. Also check whether the plan includes the feature you need. A subscription alone does not establish better learning outcomes. Until those details are verified, this article cannot identify a value-for-money winner.

Which method fits your study problem?

NotebookLM is a candidate when you want to work with your own sources. These might be lecture materials you have permission to use or your own notes. For lecture recordings, obtain permission to record and process them first. Check the currently supported input format; a transcript is not automatically accurate.

A general chatbot may help with an alternative explanation or a follow-up question, such as: “What does this definition mean in plain language?” Without specified sources, do not treat the answer as an account of your lecture. State the relevant terms and assumptions, then compare the result with your course material.

Traditional methods fit when the material is already manageable or you need to learn how to apply it. A calculation requires your own attempt at a solution. For an oral exam, practise explaining a term without an answer on screen. Generating another revision sheet does not solve that problem.

You can assess Astra AI against the same learning goal. Before paying, check how the specific app handles your material. Do not share confidential content while doing so. Without verified details, we recommend neither a subscription nor a particular plan.

What AI can do during exam preparation

A summary can help you organise a long passage. Comprehension questions can reveal where you need to reread. A definition can supply draft material for flashcards. These are working drafts, not verified subject answers or predictions of what the exam will contain.

Keep the request narrow: “Summarise this passage. Preserve the conditions attached to the definition and flag unclear points.” For questions, you could add: “Ask me about the difference between these two terms. Show the model answer only after my attempt.” These are suggested requests, not promises of a particular app feature.

Subject details are harder. An AI may omit a qualification or give a plausible but incorrect explanation. Check formulas and their assumptions against the original. If a step remains unclear, asking in a tutorial is more useful than repeatedly rephrasing the same unsupported answer.

Measure progress through your own attempt. Close the revision sheet and tackle an exercise. Afterwards, note where you got stuck without help. That gap is your next study task. The number of generated pages or audio summaries you have heard tells you little about it.

Check privacy and university rules before uploading

Do not send other people's personal data or confidential university material to an external service. Remove names and student numbers when your task does not require them. Even an apparently harmless case study may contain information about other people.

Check usage rights and university requirements for other people's notes, research documents and recordings. Access through a learning portal does not automatically permit external processing. Also establish what AI use your particular assessment allows. This article cannot replace advice from your examinations office.

Five steps from lecture material to revision notes

  1. Narrow down the material. Choose one section and put your learning goal above it. For example: “I can explain the definition and apply it to an exercise.” Keep the original with its page reference. Check that you may use the material before sharing it.
  1. Make the text readable. Extract the relevant PDF text, preferably locally. Check the reading order and look for missing formulas. A scanned page may require text recognition. Do not pass a wrongly extracted character into the next summary without checking it.
  1. Shorten a draft. Summarise the passage yourself or use a suitable summarisation aid. Simple text tools can also shorten material locally without an account. They do not replace a subject accuracy check. If you use AI, ask it to stick to the passage and check each main statement you keep.
  1. Write your own revision sheet. Arrange the checked statements under the terms used in your course. Add an example you solved yourself where useful. The Word Counter counts words and characters locally without an account. It helps you shorten text, but assesses neither understanding nor completeness.
  1. Test without the sheet and revise it. Answer your own questions or solve a suitable exercise without looking at the notes. Then compare with the original and correct errors. Keep source references and unresolved questions alongside the finished revision sheet.

This workflow does not necessarily require a new account. Extracting PDF text, counting characters and shortening a passage are separate tasks. Simple local tools can handle some of them; that does not mean every website advertising “summarisation” processes text locally. Check the particular service before entering sensitive content.

If the material needs to become a presentation, the Prepare a Presentation playbook offers a separate workflow. It replaces neither checking subject accuracy nor working with the original lecture material.

Solve one learning gap before deciding on a subscription

Choose a manageable passage you have permission to use. Check whether the result helps you explain something better or solve an exercise independently. If you just need a shorter text, start without a subscription. If you regularly work with your own sources, NotebookLM is a reasonable candidate to assess based on the documented features.

Reliable Astra AI details are missing here. That is a limit of this comparison, not a verdict against the app. Do not buy a subscription on the assumption that it has certain features. Editorial date mid-November 2026, research baseline 15 October 2026: please recheck prices and limits before buying.

Frequently asked questions

Is NotebookLM or Astra AI better for studying?
For working with your own sources, NotebookLM is the option documented here. The research baseline describes Astra AI as an app-based study assistant with a subscription model. Without verified feature and pricing details, this does not establish a reliable winner. Editorial date mid-November 2026, research baseline 15 October 2026.
Can I use NotebookLM for free?
The project facts verified on 15 October 2026 list free access with 100 notebooks, 50 sources per notebook and 50 chats daily. Editorial date mid-November 2026, with no new pricing check in November. Check the current limits before using it.
Does NotebookLM offer German audio summaries?
The research baseline from 15 October 2026 lists audio summaries in more than 80 languages, including German. Editorial date mid-November 2026; please check with the provider. An audio summary does not replace doing a practice exercise yourself.
Which Astra AI details are missing?
Specific prices, subscription terms and feature limits have not been verified. Check the actual app offer, including cancellation and privacy, yourself. This article deliberately gives no figures for those details and does not claim feature parity with NotebookLM.
May I upload lecture notes to an AI service?
That depends on the material and your university's rules. Do not upload other people's personal data or confidential documents. Obtain permission before using someone else's notes or recordings; university access is not blanket permission to send material to an external service.
Do I need an AI account for every set of revision notes?
No. You can extract text locally from a suitable PDF, shorten it yourself and count words or characters without an account. Simple local summarisation can reduce text, but does not verify subject accuracy. You still need to check the statements yourself.

No product advertising, affiliate recommendation or first-hand comparison test. Check subject content against the original and follow your university's rules.

Sources & status of these figures

Editorial date: mid-November 2026. Product details come from the project plan verified on 15 October 2026, not a fresh provider check in November. Check prices and limits before buying. No verified pricing or feature details are available here for Astra AI.

  • Studio-tools project plan, AI tools for studying section: facts verified on 15 October 2026 for NotebookLM, also called Gemini Notebook; Astra AI classified only as an app-based study assistant with a subscription model.
  • Topic prompted by the gutefrage question Astra AI or NotebookLM? dated 15 October 2026. The question is not evidence of product performance or learning outcomes.