Google's July 2026 AI updates deliver three practical tools for businesses: AI avatars in Google Vids that put your face and voice on camera without filming, Lyria 3 Pro for generating full 3-minute songs from text, and the rebrand of NotebookLM to Gemini Notebook with built-in code execution. All three shipped in a single week and are available now to paid Google AI subscribers. Here's what each one does, what it costs, and how to actually use them for work.
Last verified: August 1, 2026
- Google Vids personal avatars: live since July 16, 2026, 7M+ monthly users
- Lyria 3 Pro: 3-minute AI songs, available across 6 Google platforms
- NotebookLM → Gemini Notebook: renamed July 16, 2026, 30M+ users
- Pricing/availability changes fast — verify at Google's official pages before relying on details below.
What Are the New Google Gemini AI Updates for July 2026?
The three updates are: (1) Gemini Omni personal avatars in Google Vids, which let you create a digital clone from one selfie and one voice sample; (2) Lyria 3 Pro, Google's AI music model that generates full songs up to 3 minutes with verse-chorus-bridge structure; and (3) the Gemini Notebook rebrand (formerly NotebookLM), which adds a secure cloud computer for running code inside your research notebooks. All three were announced on or around July 16, 2026.
Google is stitching these into one ecosystem: writing lives in Gemini, research in Gemini Notebook, video in Vids powered by Gemini Omni, and music in Lyria — now built into the Gemini app and Vids as well. The same Gemini Omni model that powers avatars has already spread to YouTube Shorts and Google Ads, and Gemini Notebook is planned for integration into Google Search's AI mode.
What Is Google Vids Personal Avatars and How Do They Work?
Google Vids personal avatars let you upload one selfie and a short voice recording, then type a script — your digital avatar reads it on camera without you ever filming. The feature launched July 16, 2026 as part of two updates to Google Vids: Gemini Omni text-to-video generation and personal avatars.
Here's how to create one:
- Open Google Vids at vids.new (requires a Google AI Pro, AI Ultra, or Google Workspace business account).
- Upload a selfie — a single photo of your face, well-lit, looking at the camera.
- Record a short voice clip — a few seconds of your natural speaking voice. This becomes the avatar's vocal fingerprint.
- Type your script — write what you want the avatar to say, the same way you'd type a document.
- Generate the video — your avatar delivers the script on screen, with your face and voice, no camera required.
- Edit with natural language — use Gemini Omni to swap backgrounds, adjust lighting, or fix a line without re-recording the whole clip.
The avatar is locked to your Google account — nobody else can create an avatar of you from your photo. Access is restricted to users 18 and older, and the feature is not available in the European Economic Area, Switzerland, or the United Kingdom at launch.
Every clip generated with Omni carries an invisible SynthID watermark embedded in every video frame, so anyone can verify the video was AI-generated. This watermark survives compression, cropping, and common edits.
Who can use Google Vids personal avatars?
| Plan | Price | Vids Avatar Access |
|---|---|---|
| Google AI Plus | $7.99/mo | 10 free generations/mo (avatar + Veo combined) |
| Google AI Pro | $19.99/mo | Full access |
| Google AI Ultra ($100 tier) | $100/mo | Full access + higher usage limits |
| Google AI Ultra ($200 tier) | $200/mo | Full access + maximum limits |
| Google Workspace Business | Varies | Full access |
Sources: Google Blog (July 16, 2026), Google AI plans page, 9to5Google (June 17, 2026 — free tier details).
How Does Gemini Omni Work for Video Editing?
Gemini Omni is Google's multimodal video generation and editing model, first shown at Google I/O in May 2026. It handles three jobs that usually take three separate people: writing the script, filming the footage, and editing the final cut. You give it a text prompt and optionally a reference photo, and it generates a clip from both inputs.
The editing is where it stands out. Instead of regenerating from scratch when something is wrong, Omni supports step-by-step edits — you describe the change in plain language ("swap the background to a cityscape," "make the lighting warmer") and Omni edits just that part while preserving the rest of the clip. This is closer to how a video editor actually works than the generate-and-pray approach of earlier AI video tools.
Within about two months of its I/O debut, Gemini Omni had already been integrated into YouTube Shorts (for AI-powered remixing), Google Ads Asset Studio, and Google Vids. That's the same model running across three completely different products — Google isn't building separate AI for each surface; it's wiring one brain into everything.
What Is Google Lyria 3 Pro and How Is It Different from Lyria 3?
Lyria 3 Pro generates full AI music tracks up to 3 minutes long with structural awareness — meaning it understands intros, verses, choruses, and bridges, and can follow instructions like "start soft with piano, build into a big chorus, then slow back down." The base Lyria 3 model (launched February 2026) only handles 30-second clips — a jingle, not a song.
Lyria 3 vs. Lyria 3 Pro comparison
| Feature | Lyria 3 (Feb 2026) | Lyria 3 Pro (Mar 25, 2026) |
|---|---|---|
| Max track length | 30 seconds | 3 minutes |
| Structural control | Basic | Intros, verses, choruses, bridges |
| Platform availability | Gemini app, AI Studio | Gemini app, Vids, Vertex AI, Gemini API, AI Studio, ProducerAI |
| Enterprise API | Not available | Public preview on Vertex AI |
| SynthID watermark | Yes | Yes (survives audio edits) |
| Image-to-music input | No | Yes |
Where can you access Lyria 3 Pro?
Lyria 3 Pro is available across six platforms simultaneously — the widest distribution Google has ever given an AI music model:
- Gemini app — for paid subscribers (AI Plus, Pro, Ultra)
- Google Vids — for Workspace and AI Pro/Ultra customers
- Vertex AI — public preview for enterprise developers
- Gemini API — paid preview for developers
- Google AI Studio — for testing and prototyping
- ProducerAI — for professional music creators
The key business advantage over competitors like Suno and Udio: Google integrated Lyria 3 Pro directly into Google Workspace. If your team already uses Google Docs, Sheets, and Slides, the same ecosystem now handles your background music — no separate tool or licensing fee. All outputs carry a SynthID watermark, and Google states the training data is content they have legal rights to use, which matters if your legal team asks about copyright.
How to Generate AI Music with Lyria 3 Pro: A Practical Workflow
Here's a real workflow for creating background music for a business video without touching a stock music site:
- Open the Gemini app or Google Vids and navigate to the music generation panel.
- Describe the mood and structure: "Create a 2-minute upbeat track that feels confident and motivating, built for a business audience. Start with a soft piano intro, build into an energetic chorus with light drums, and keep it clean enough to sit under a voice-over without sounding distracting."
- Generate the track — Lyria 3 Pro produces a stereo audio file with the structure you described.
- Export or embed — in Google Vids, the track can sit directly behind your avatar video. Via Vertex AI, you get the audio file for use in any editor.
- The SynthID watermark is embedded automatically — no extra step needed, and it survives format conversion.
For longer or more complex pieces, specify the full structure in your prompt: "Start with a 15-second piano intro, move into verse 1 (30 seconds), chorus (30 seconds), verse 2 (20 seconds), bridge (20 seconds), final chorus (25 seconds)." Lyria 3 Pro follows this like a session musician reading a chart.
What Happened to NotebookLM? Why Was It Renamed to Gemini Notebook?
Google renamed NotebookLM to Gemini Notebook on July 16, 2026 because "LM" (language model) meant nothing to most users. The rebrand ties the product to Google's unified Gemini brand — same blue-purple gradient logo as the rest of the AI lineup. Josh Woodward, VP of Google Labs, announced the change.
Nothing breaks: your existing notebooks carry over, shared links redirect automatically, and the free tier stays free. Two things actually changed beyond the name:
- Secure cloud computer: Every notebook now has a sandboxed code execution environment. Gemini Notebook can write and run Python code to clean data, perform calculations, and generate charts from your sources — without you leaving the notebook or setting up a separate development environment.
- Cross-Google integration: Notebooks now sync with the Gemini app for chat-based organization, and Google plans to surface notebooks inside Search's AI mode.
Gemini Notebook user count
Before the rename, NotebookLM had over 30 million users and 600,000 organizations using it. That scale is why the rebrand matters — it's not a niche experiment but a tool millions of people built workflows around.
Gemini Notebook vs. Gemini: What's the difference?
Gemini Notebook stays a standalone research product — you upload your own documents, papers, and notes, and it answers questions using only your sources (not the whole internet). The Gemini app (chat) is for general-purpose AI conversation. Think of the relationship like this: Gemini is the brain you talk to; Gemini Notebook is the brain you point at a specific pile of information and say "analyze just this."
How to Use Gemini Notebook for Content Research: A Practical Workflow
Here's a workflow for turning customer questions into content ideas using Gemini Notebook's new capabilities:
- Create a notebook at notebooklm.google.com and upload your source documents — call transcripts, customer support tickets, forum posts, survey results.
- Ask structured questions: "Look through these call transcripts and posts from customers. Find the three questions they ask most often about [your topic]. For each one, summarize the question in one sentence and suggest one tutorial topic that would answer it clearly for a beginner."
- Use the cloud computer for analysis: Ask Gemini Notebook to run code that counts question frequency, sorts by category, or cross-references with your existing content to find gaps.
- Export the results — you get ready-made tutorial ideas grounded in what your actual audience is asking, not guesses.
This is the practical leap: before the code execution feature, Notebook was a reading and summarizing tool. Now it's a lightweight analytics environment that can crunch the data you upload.
What Does This Mean for Small Businesses and Content Creators?
If you produce any kind of video, audio, or research content on a schedule, these three updates collapse separate tools and workflows into one Google subscription:
- Video: Instead of filming yourself, write the script and let an avatar handle the delivery — useful for onboarding videos, weekly updates, and training content. Update the words, regenerate, done.
- Music: Instead of licensing stock tracks, generate custom background music that matches your video's tone — with a SynthID watermark so there's no copyright ambiguity.
- Research: Instead of switching between a note-taking app and a code editor, run analysis directly inside the notebook where your sources live.
The catch: all three require a paid Google AI plan. Google AI Plus at $7.99/month gets you 10 generations (shared between avatars and Veo video). AI Pro at $19.99/month gives you full access with reasonable limits. The Ultra tiers ($100 or $200/month) add higher limits, Project Mariner, and more. If your team is already on Google Workspace, the avatars come included with business plans.
One tradeoff worth naming: an avatar still reads as an avatar. Use it for utility content — onboarding, product walkthroughs, policy videos — where the information is the point. Use your real face for anything that trades on personal connection.
Google's Convergence Strategy: Why All Three Updates Shipped in One Week
These aren't three random features. They're pieces of the same machine:
| Capability | Tool | Where it lives |
|---|---|---|
| Writing | Gemini | Gemini app, Workspace |
| Research | Gemini Notebook | notebooklm.google.com |
| Video | Google Vids + Gemini Omni | vids.new |
| Music | Lyria 3 Pro | Gemini app, Vids, Vertex AI |
| Search | Coming soon | Google Search AI mode |
Before, each was a separate product. Now Google is connecting them — the same avatar tool that makes your video can also score the music behind it, and the research tool feeds into the content you create. If you produce content on a schedule, the value isn't any single feature — it's that you can write the script, research the topic, create the video, and generate the soundtrack all inside one ecosystem, without switching subscriptions.
The pace tells you what's coming next. Gemini Omni went from an I/O demo to three shipping products in two months. Lyria went from 30-second jingles to 3-minute songs in one month. Gemini Notebook is moving from a standalone tool into Search itself. If this trajectory holds, expect Omni and Lyria inside more Google products before the end of 2026.
For builders and developers, these are also available as APIs — Lyria 3 Pro via Vertex AI and the Gemini API, Gemini Omni via Google Vids and the broader Gemini platform. If you're building tools that sit on top of AI-generated video or audio, Google's API access means you can integrate these capabilities into your own products. For more on orchestrating AI agents that can chain these tools together, see our guide to orchestrating AI agents like a company.
FAQ
Q: How much does it cost to use Google Vids personal avatars?
A: Personal avatars are available to Google AI Pro ($19.99/month), AI Ultra ($100 or $200/month), and Google Workspace business customers. Free personal accounts get 10 video generations per month, shared between avatars and Veo video generation, as of June 2026.
Q: Can someone create an avatar of me without my permission?
A: No. Personal avatars are locked to your individual Google account and restricted to the account holder's own likeness. Nobody can generate an avatar of someone else from a stray photo. Access is limited to users 18 and older.
Q: What is the SynthID watermark and does it affect video quality?
A: SynthID is Google DeepMind's invisible digital watermark embedded directly into AI-generated video, audio, text, and images. It is imperceptible to humans, does not degrade quality, and survives compression, cropping, and common edits. It exists so anyone can verify whether content was AI-generated — you can upload a clip to Gemini and ask if it carries a SynthID watermark.
Q: Is Lyria 3 Pro music free to use commercially?
A: Google states Lyria 3 Pro was trained on content Google has the legal right to use under its terms of service and partner agreements. Google's current terms of service for AI-generated content should be reviewed for specific commercial licensing rights, particularly for enterprise use via Vertex AI. Unlike Suno and Udio, which settled copyright lawsuits in 2025, Google's training data legal position is designed to be defensible for enterprise deployment.
Q: What's the difference between Lyria 3 and Lyria 3 Pro?
A: Lyria 3 (launched February 2026) generates clips up to 30 seconds with basic style controls. Lyria 3 Pro (launched March 25, 2026) generates full tracks up to 3 minutes with structural composition control — you can specify intros, verses, choruses, and bridges — plus image-to-music input and time-aligned lyrics.
Q: Do my existing NotebookLM notebooks still work after the rename?
A: Yes. Google confirmed all notebooks carry over untouched, shared links redirect automatically, and the free tier continues unchanged. The rename is cosmetic — the same product now has the Gemini brand. The one new capability is the secure cloud computer for running code inside each notebook.
Q: Is Gemini Notebook different from the Gemini chat app?
A: Yes. Gemini Notebook is a standalone research tool where you upload your own documents and get answers grounded only in your sources. The Gemini chat app is for general-purpose AI conversation using the entire internet. You can read our deeper analysis of AI content quality frameworks for how to evaluate what each surfaces.
Discussion
0 comments