NotebookLM and Google Gemini’s New Features Are Insane

Futuristic illustration of a notebook transforming into short video frames with glowing network nodes representing AI connections and automation features, no text

Short Videos, MCP Connections, Spark Automations, and More

Google just dropped a wave of NotebookLM and Gemini updates that seriously change what these tools can do. The biggest headline is easy to spot: NotebookLM can now create short form video content, and Google Gemini can now connect to custom apps through MCP servers. But that is only the beginning.

There are also major improvements to flashcards, prompt iteration, real time Spark automations, connected app workflows, model selection for image generation, and even avatar based video creation. If you use AI for content creation, research, business systems, or automation, these updates are a big deal.

This article breaks down what changed, why it matters, and how to actually use these new features in a practical way.

Meta description: Explore NotebookLM and Google Gemini’s latest features, including short video creation, MCP app connections, Spark automations, avatars, and more.

Suggested category: AI Tools

Suggested tags: NotebookLM, Google Gemini, Gemini Spark, MCP, AI automation, short form video, AI content creation, AI avatars

Table of Contents

NotebookLM can now create short form videos

This is one of the most surprising and useful updates in the whole release.

Inside NotebookLM, there is now a Video Overview option that lets you generate short form content from your source material. That means if you already have notes, documents, research, or long form content inside a notebook, you can turn it into clips built around a specific angle or topic.

For example, if your notebook contains material about making money by buying cars, you can generate shorts around focused ideas like:

  • Smart asset investing
  • Perfect car aesthetics
  • Top five cars to invest in in 2026

That matters because the tool is not forcing you into one generic output. You can point it toward a very specific concept and get multiple short video assets from the same knowledge base.

Why this update is such a big deal

For content creators, this is an obvious win. For businesses, it is arguably even bigger.

You can now take existing educational material, internal knowledge, product notes, market research, or long form media and quickly turn it into social ready short form pieces. That reduces the gap between research and distribution.

It also creates a free or low cost workflow for slicing larger ideas into shorter content units without needing a separate clipping tool.

If you select a long form piece inside NotebookLM, you can effectively use this as a shortcut to break that longer asset into smaller short form segments. That is a huge upgrade if you want to repurpose content efficiently.

Be specific with custom prompts

There is one important catch.

If you use a custom topic, you need to be as specific as possible. If your prompt is vague, the result will probably be broad, generic, and less aligned with what you actually wanted.

That is true across most AI tools, but here it really matters because short form content lives or dies on clarity. A focused prompt gives you a focused hook, a clearer angle, and a more usable output.

A good approach is to define:

  • The exact topic
  • The intended angle
  • The audience or use case
  • Any timeframe or ranking format

So instead of asking for a short about cars, ask for something like top collector cars with appreciation potential in 2026. That level of precision tends to produce much stronger material.

What the generated videos include

Once the short is ready, NotebookLM gives you several useful controls. You can:

  • Play the generated video
  • Download it
  • Share it
  • Delete it
  • View the prompt behind it
  • View the source material used
  • Iterate on the output

Generation is not instant. In testing, it took roughly two to three minutes, especially when creating multiple outputs at once. That is not bad considering it is producing a complete short form asset.

The really useful part is the transparency. Being able to inspect the prompt and source makes it much easier to improve future generations and maintain control over the final output.

NotebookLM now makes iteration much better

One of the smartest upgrades here is how NotebookLM handles iteration.

When you generate something and click Iterate, the system automatically pulls in the original prompt and the settings you used. Then you can modify them rather than starting over from scratch.

That sounds small, but it solves a very common AI workflow problem.

Normally, when an output is close but not quite right, you have to reconstruct what you asked for, remember your settings, and manually rewrite everything. Now the adjustment loop is much faster.

This is especially helpful for:

  • Refining short videos
  • Improving educational outputs
  • Testing alternate angles
  • Making format changes without losing the original direction

Flashcards in NotebookLM are finally customizable

Another very practical update is improved control over flashcards.

Previously, flashcards were more rigid. Now you can customize them based on the kind of study or review experience you want.

You can adjust:

  • Number of cards such as fewer, standard, or more
  • Difficulty level such as easy, medium, or hard
  • Topic focus based on what you want the cards to cover

If you use NotebookLM for learning, teaching, research breakdowns, or knowledge retention, this gives you much more control over the final study material.

It also fits into the broader theme of these updates: Google is slowly turning these tools from one shot generators into systems you can actually shape.

Gemini Spark can now connect to custom apps through MCP

This is probably the most important Gemini update in the release.

Inside Gemini Spark, under connected apps, there is now support for custom apps. In practical terms, this means you can connect Gemini to anything that exposes an MCP server.

If you have been waiting for Gemini to catch up in the automation and integration game, this is a huge step.

What MCP support really means

MCP support means Gemini is no longer limited to a closed set of built in integrations. You can now plug it into external systems and workflows in a much more flexible way.

That opens the door to:

  • Custom business workflows
  • Third party app automation
  • Cross platform task execution
  • AI assistants that can take action across your stack

And if you use a platform like Zapier, the scale jumps fast.

Using Zapier as an MCP bridge

One of the easiest ways to make this useful is to connect Gemini Spark through Zapier. By creating a new MCP server in Zapier and adding tools there, you can give Gemini access to thousands of apps through a single connection path.

The rough process looks like this:

  1. Create a new MCP server in Zapier
  2. Select the custom or other server option
  3. Add the tools or apps you want available
  4. Copy the connection details
  5. Paste them into Gemini Spark’s custom app connection flow
  6. Use any required token, client ID, and client secret

Once that is connected, Gemini becomes dramatically more useful. Instead of only generating ideas, it can help coordinate actions across external platforms.

If you want a broader understanding of automation ecosystems, it is worth reviewing resources like Zapier and Google’s own Gemini product documentation as these features continue to evolve.

Gemini can now monitor the web and react in real time

This is where Spark starts getting genuinely interesting.

Gemini Spark can now monitor topics across the web, including areas like:

  • News
  • Finance
  • Sports
  • Local events
  • Other web based updates

It can then run scheduled tasks or trigger workflows based on specific conditions.

One example is setting up a search based schedule for local food pop ups. You can tell Gemini to monitor for announcements in your city, email the details, and even suggest calendar time to check them out.

That is a simple example, but the pattern is what matters. Gemini is moving from passive assistant to active monitoring system.

How to set up schedules in Gemini

If you want to use this yourself, the setup is straightforward:

  1. Open Schedules in Gemini
  2. Choose Create with Gemini
  3. Describe what you want monitored or repeated
  4. Define the action to take when the condition is met

Gemini can work across your:

  • Gmail
  • Calendar
  • Drive
  • Web activity

That means you can build recurring routines or event based automations like:

  • A morning news digest
  • Email based drafting workflows
  • Alerts when specific topics appear online
  • Calendar actions tied to incoming information

And once MCP connected apps are layered on top, the possible workflows expand even more.

Gemini Spark is coming to desktop and mobile

Spark is also expanding beyond one interface.

It is being rolled out inside the Gemini desktop app, and similar functionality is also making its way to phones. That matters because it pushes Gemini toward a much more accessible operating model where your AI workflows are available across devices.

The bigger picture here is obvious: Google is building toward an assistant that can control a cloud based work environment connected to your apps, tasks, and automations.

That starts to resemble the direction other leading AI platforms have been pushing toward, but now Gemini is finally making moves that feel competitive.

Skills are becoming essential inside Gemini

If you are using connected apps, schedules, or Spark at all, you should absolutely be setting up skills.

Skills are basically custom reusable instructions that Gemini can apply automatically when relevant, or manually when you invoke them with a slash command.

This is important because it reduces repetitive prompting.

Instead of re explaining the same preferences or workflow every time, you can create a skill once and let Gemini use it as needed.

Why skills matter

Skills help with consistency and speed. They can be used to standardize:

  • Writing style
  • Task preferences
  • Formatting rules
  • Business processes
  • Response structures

Gemini also includes a dashboard style view where you can filter Spark activity by status, including recent, scheduled, needs input, in progress, and completed. That makes the system feel more like an operations layer rather than a single prompt box.

Model choice now matters more for image generation

Another easy to miss update is how Gemini handles image generation depending on which model you are using.

If you generate an image while using Flash Lite, Flash, or Pro, the experience may differ because each option uses a different Gemini model.

This can affect:

  • Image quality
  • Edit reliability
  • Generation speed
  • Output consistency

If your image edits look wrong, feel slow, or come out lower quality than expected, the issue may not be your prompt. It may simply be the model choice.

That is an important operational detail because many people assume all image generation modes work the same way. They do not.

The Gemini “plus” menu is turning into a full creative toolbox

Another update worth noting is how much functionality now sits behind the plus button inside Gemini.

From there, you can access a broad set of tools and modes, including:

  • Deep research
  • Guided learning
  • Personal intelligence controls
  • Canvas
  • Music creation
  • Video creation
  • Image creation
  • File uploads
  • Drive imports
  • Photo inputs
  • Code imports
  • Notebook connections

This really shows the direction of the platform. Gemini is not just a chatbot anymore. It is becoming a central workspace for generation, research, automation, and media production.

If your site covers related workflows, this is a perfect place to add internal links to articles about AI image prompts, AI workflow design, or Google productivity tools.

Gemini avatars may be one of the most underrated features

You can now create an avatar inside Gemini and use it in video creation.

In practical terms, this gives you a digital clone that can be used for content, intro videos, team communication, and outreach style workflows.

Some obvious use cases include:

  • Creating branded content quickly
  • Sending short internal team updates
  • Producing video intros without filming each time
  • Testing personalized outreach formats

What stands out is how realistic these avatars can be. In many cases, they compete well with dedicated avatar tools, and the process to remake one is very fast. If you do not like the result, creating a new version only takes about 30 seconds.

That low friction makes experimentation much easier.

Suggested media to include in this article

  • An image of NotebookLM’s Video Overview interface with alt text: NotebookLM short form video generation tool
  • An image of Gemini Spark custom app connection settings with alt text: Google Gemini Spark MCP custom app connection
  • An infographic showing how schedules, skills, and MCP integrations work together with alt text: Gemini Spark automation workflow with schedules and MCP

Why these updates matter right now

There is a bigger pattern across all of this.

Google is making NotebookLM and Gemini more actionable. NotebookLM is becoming more useful for repurposing and packaging knowledge. Gemini is becoming more capable as an automation and execution layer.

Together, these updates point toward a workflow where you can:

  1. Research or collect source material
  2. Turn that material into short content or study assets
  3. Connect your AI assistant to external apps
  4. Set ongoing monitoring and scheduling rules
  5. Produce creative assets including videos, images, and avatars

That is a much more complete stack than what Google had before.

And honestly, the MCP support alone makes Gemini feel much closer to where it needed to be. For a long time, that gap was obvious. Now it feels like Google is finally starting to close it.

FAQ

Can NotebookLM really create short form videos now?

Yes. NotebookLM now includes a Video Overview feature that can generate short form content based on the source material in your notebook. You can guide it with custom topics to create more focused outputs.

What is the best way to get better short video results in NotebookLM?

Use highly specific prompts. Broad prompts tend to produce generic content. Define the angle, topic, and intended outcome as clearly as possible.

What does MCP support in Gemini Spark allow you to do?

It allows Gemini to connect to custom apps and external systems through MCP servers. This makes Gemini far more flexible for automation and connected workflows.

Can Gemini Spark monitor the web for updates?

Yes. Gemini Spark can monitor topics across the web, including news and local events, and then trigger scheduled actions such as email alerts or calendar suggestions.

Why do Gemini image generation results vary so much?

Different modes such as Flash Lite, Flash, and Pro use different Gemini models. That can change speed, quality, and editing behavior, so model selection matters.

Are Gemini avatars useful for business content?

Yes. They can be used for team updates, intros, content creation, and communication workflows. The setup is fast, and creating a replacement avatar is simple if you want a better result.

Final thoughts

These NotebookLM and Google Gemini updates are not small quality of life tweaks. They meaningfully expand what both products can do.

NotebookLM is becoming a practical engine for content repurposing and learning assets. Gemini is becoming a more serious platform for automation, connected actions, and creative production.

If you are building AI workflows right now, these are exactly the kinds of updates worth testing early because they can change how you structure your entire process.

If you want to keep exploring this space, consider linking this piece to related guides on prompt engineering, AI automations, or content repurposing workflows. And if you found this useful, share it with someone building with Gemini or NotebookLM right now.

Facebook-f Twitter Pinterest

Leave a Reply

Your email address will not be published. Required fields are marked *

Most Read

Subscribe To Our Magazine

Download Our Magazine