Google Gemini’s New Changes Are Mind Blowing (Lyra 3 Music + Scheduled Actions + New Gems)

Google just shipped a stack of updates that make Gemini feel less like a chatbot and more like an actual creative and productivity system. The big headline is Lyra 3, a new music generation upgrade that can produce tracks up to three minutes long. On the workflow side, Gemini now supports scheduled actions across apps, so you can set it up once and let it handle repetitive tasks for you.

Then there is a whole layer of “how do I build with this?” improvements, including a new Gems from labs experience for creating mini-app style workflows, plus Chrome features like Ask Gemini and auto-browse for more hands-on browser assistance.

This guide breaks down what changed, how to use each feature, and the real-world use cases these updates unlock.

Table of Contents

1) Lyra 3: Generate music up to three minutes (fast, thinking, and pro)

The most exciting addition is Lyra 3, Google’s latest music generation model inside Gemini. When you hit Create music, you can choose different generation modes:

  • Fast: produces music around 30 seconds
  • Pro: generates music up to three minutes
  • Thinking: sits in between, typically prioritizing more deliberate output than fast

That alone matters because longer tracks are useful for more than just background. Three minutes is enough for short intros, podcast segments, reel-style edits, and “loopable” themes you can remix or structure into bigger compositions.

Describe a track, or pick a mood and let Lyra 3 do the work

Lyra 3 supports prompt-based creation. You can describe a track directly, or tell it the feeling or mood you want.

One example: a “team anthem” prompt for a specific team. If you request something like a “Team Anthem for the New York Knicks for the playoffs,” the model interprets the task details, builds parameters, and generates a modern melodic trap anthem style.

What is especially interesting is the localization effect. A prompt tied to New York imagery tends to produce a different “vibe” than a prompt tied to another region or identity. In other words, your creative intent is not just generic style transfer. It is treated as part of the musical context.

Examples of what you can generate

Here are practical ways people can use Lyra 3 right now:

  • Background music for YouTube, TikTok, Shorts, and podcasts
  • Content intros (music stings, theme beds, channel bumpers)
  • Remix starting points for producers and editors
  • Sports hype themes for highlight packages and team content
  • Brand mood tracks for ads, landing pages, and product videos

If you create content regularly, this becomes a “music assembly line.” You can iterate quickly by changing just one piece of the prompt: the mood, the tempo vibe, the genre direction, or the target audience.

2) Scheduled Actions: automate multi-app workflows inside Gemini

Lyra 3 is the creative headline, but scheduled actions are where Gemini starts to feel like a personal operations layer.

The new idea is simple: set up triggers and schedules so Gemini can run tasks for you at specific times, instead of you doing everything manually each day.

Start with Gemini presets

Gemini includes preset templates that make onboarding painless. For instance:

  • Morning motivation
  • Latest news updates for a region

Presets matter because they reduce friction. You can take something working immediately and then customize it to your needs.

Connect your apps and let Gemini execute

To make scheduled actions actually useful, you link your connections (like Gmail). Then Gemini can carry out tasks. A typical setup might look like:

  • Create an email brief
  • Ask Gemini to check Gmail for important messages needing replies
  • Schedule it for specific times
  • Receive a structured summary so you can act faster

This is valuable for anyone who gets overwhelmed by incoming tasks. You stop relying on “I will check later” and start relying on a system that brings information to you.

Build multiple schedules for different roles

Once you get the hang of it, you can set up multiple scheduled actions depending on your workflow. For example:

  • Content inspiration prompts
  • Sponsorship email checks
  • Regional news briefs
  • Task lists pulled from email threads

In practice, scheduled actions help you separate “thinking time” from “execution time.” Gemini does the repetitive scanning and formatting so you can spend your brainpower on decisions.

3) MyPromptBuddy Chrome extension: shortcuts + prompt optimization

Not every upgrade lives inside Gemini. One tool highlighted alongside these updates is MyPromptBuddy, a Chrome extension that helps you use better prompts more consistently.

Why prompts matter more than people think

If you have ever gotten a “meh” answer from an LLM, the reason is often not the model. It is the input. Even small changes to prompt structure can dramatically change the output quality.

Two key features: shortcuts and a prompt optimizer

MyPromptBuddy focuses on practical prompt workflows:

  • Shortcuts: store prompts so you can insert them instantly without losing context in a long chat thread
  • Prompt optimizer: transform a basic prompt into a more professional, structured one

There is an important workflow benefit here. Instead of running “prompt trial and error” back and forth, you can upgrade your prompt once and reuse it.

Example: improved DCF model prompts

In a DCF valuation example, a basic prompt can produce an “okay” answer. But an optimized prompt adds structure and outputs more complete information, including things like:

  • Terminal value
  • Valuation summary
  • Applied price
  • Exportable breakdowns (like a sheet-friendly structure)

That is the difference between “it tried” and “it produced something you can actually use.”

4) New Gems from labs: build mini apps with images and multi-step logic

Another major shift is the emergence of new Gems from labs. Instead of only using Gemini as a single chat, you can create mini-app style tools powered by Gemini and underlying agents.

What is a “Gem” in practical terms?

A Gem is essentially a packaged workflow. You can:

  • Create a tool once
  • Reuse it repeatedly
  • Share it so other people can use the same experience

Example Gem: fridge leftovers to personalized recipes

The demo-style example is a recipe workflow that starts with what is already in your fridge. The concept:

  1. Take a photo of leftovers
  2. Upload it inside the Gem
  3. Generate multiple recipe ideas based on what is in the image
  4. Provide step-by-step cooking instructions and ingredient guidance

After generation, the interface includes an advanced editor. That is where the power shows up for builders: you can modify the workflow, adjust steps, and control which models get used.

Go beyond generic recipes: add granular model control

In the advanced editor, one highlighted upgrade is the ability to configure how models are selected. Instead of using an agent that “uses every model,” you can choose specific models for more granular control.

This matters if you are trying to balance:

  • Quality (more capable models for tricky steps)
  • Speed (lighter models for simpler steps)
  • Cost (more efficient routing)

If you are building for real users, this kind of control turns Gemini into an engineering platform rather than a one-off assistant.

Suggested multimedia: Add screenshots of the “new gem” creation flow and the advanced editor steps. Use alt text like “Gem from labs advanced editor showing recipe generation steps.”

5) Ask Gemini in Chrome: summaries, comparisons, and browser automation (auto-browse)

Finally, there is the browser layer. On Google Chrome, the interface now includes an Ask Gemini option in the right-hand corner.

This is a big deal because it changes where Gemini can help. Instead of copying and pasting content into a chat, you can ask questions directly while you are already on a page.

Common tasks Ask Gemini supports

Based on the described functionality, Ask Gemini can help with:

  • Instant summaries of what is on the page
  • Comparing options (for example, evaluating multiple products or cars based on requirements)
  • Assistance with customer support style questions
  • Interacting with apps and workflows while you browse

Share tabs for deeper help

There is also a “plus” action that lets you choose which tabs Gemini should share or access. This is how you get deeper context without manually juggling windows.

Auto-browse: Gemini takes over your browser (subscription caveat)

One of the most powerful features is auto-browse. The idea is that Gemini can take control of your browser and do tasks you would normally do manually.

Important transparency: auto-browse is described as available for Google AI Pro and Ultra subscribers. If you have those plans, you can use the feature to automate actions across sites and multi-step browsing tasks.

External reference: For the latest official details and plan eligibility, check Google’s Gemini documentation: https://support.google.com/gemini/.

Suggested multimedia: Include a short GIF or diagram showing “Ask Gemini summary” and “Auto-browse task flow.”

Putting it all together: real use cases for 2026-style workflows

Individually, these features are cool. Together, they can form a full creative and productivity loop.

Use case 1: content pipeline for creators

  • Lyra 3 generates a background track matching your video mood
  • Gemini scheduled actions pulls inspiration and drafts briefs at set times
  • MyPromptBuddy helps you standardize prompts for scripts, hooks, and shot lists
  • Chrome Ask Gemini summarizes research pages so you can write faster

Use case 2: small team operations and inbox sanity

  • Scheduled actions scan and summarize key emails
  • Chrome helps compare options across vendors
  • Gems from labs can package internal workflows into shareable mini tools

Use case 3: building lightweight “apps” with images

  • Use a Gem to turn uploaded photos into structured outputs (like recipes)
  • Customize the advanced editor to control model choices
  • Share the Gem so others can reuse the workflow

That is the direction: less one-time chat, more reusable systems.

Internal links suggestion: Add links to your site pages like “How to write better prompts,” “Gemini productivity workflows,” and “LLM automation best practices.” (If you share your site URL, I can draft the exact anchor text and target URLs.)

FAQ

How long can Lyra 3 generate music in Gemini?

In pro mode, Lyra 3 can generate music up to three minutes. Fast mode produces around 30 seconds.

Do scheduled actions require connecting apps like Gmail?

Yes. The feature becomes truly useful after linking your connections so Gemini can take actions, like checking for important emails and creating an email brief on a schedule.

What can Gems from labs do?

Gems from labs let you create mini-app style workflows, including image-based tools. You can also use an advanced editor to modify steps and control underlying model usage, then share the Gem with others.

What is the Ask Gemini feature in Chrome?

Ask Gemini adds an interface in Chrome for summaries, comparisons, and assistance while you browse. It can also support tab sharing for more context and, for eligible plans, auto-browse automation.

Is auto-browse available for everyone?

No. Auto-browse is described as available for Google AI Pro and Ultra subscribers.

Why use MyPromptBuddy instead of writing prompts manually?

It helps you reuse high-quality prompts via shortcuts and upgrade weaker prompts using a prompt optimizer, reducing back and forth and improving output consistency.

Next step: choose one feature and build a repeatable workflow

Here is the simplest path forward. Pick the one change that solves a real problem in your day:

  • If you need faster creative work, start with Lyra 3.
  • If your inbox or research routine is draining, set up scheduled actions.
  • If your prompts are inconsistent, install MyPromptBuddy and optimize your core templates.
  • If you want to package workflows for others, try a Gem from labs.
  • If you want help directly in your browser, use Ask Gemini.

CTA: Leave a comment with the first workflow you plan to automate or the music style you want Lyra 3 to generate. If you found this useful, share it with a creator, operator, or builder who loves practical AI tools.

Suggested categories/tags: Google Gemini, Lyra 3, AI music generation, Gemini scheduled actions, Chrome Ask Gemini, auto-browse, LLM prompts, MyPromptBuddy, Gems from labs, productivity automation.

This article was created from the video Google Gemini’s New Changes Are Mind Blowing 👀 (Lyria 3 Update) with the help of AI.

Share this post