Google Labs is Google's home for AI experiments: one page where you can try early tools such as Stitch, Opal, Jules and Flow Music before they become everyday products. As of Nov 2026 every experiment carries a label, Try It Now, Learn More or Join Waitlist, and that label is the fastest way to tell what you can use today.
The Labs page calls itself exactly that, Google's home for AI experiments, and its pitch is short: be the first to experiment. Google says you can get early access, share your feedback and help turn these technologies into the products you use every day. There's a Labs newsletter for news of new experiments and a trusted tester programme for people who want to test what Labs builds next.
Products that started life as Labs experiments
This is the part most people don't know. Several of Google's best-known AI products began here under a different name. The Labs page has a section called Life beyond the Lab that lists them, and says some experiments outgrow the lab and become part of everyday life:
| Name in the lab | Name now | How Labs describes it |
|---|---|---|
| Bard | Google Gemini | Your personal, proactive AI assistant for writing, planning and brainstorming |
| VideoFX | Google Flow | Your AI creative studio for creating and refining your stories |
| Project Tailwind | Gemini Notebook | Your research and thinking partner, grounded in information you trust |
| ProducerAI | Google Flow Music | Your music studio for original songs and music videos with an AI producer |
| MakerSuite | Google AI Studio | Your vibe coding platform for building production-ready apps and websites |
The practical lesson: if you're searching for a tool by its old name, you'll find old advice. Search the new name. And notice the pattern. The tools that graduated are the ones lots of people used and talked about, which is why your feedback inside an experiment carries weight.
The Labs lineup as of Nov 2026
Here's everything on the Labs page when I checked, with Google's own one-line description and the label on each tile. The right-hand column is my view of who should look first.
| Experiment | What Labs says it does | Label | Who I'd point at it |
|---|---|---|---|
| Google Flow Music | Create full songs with an AI music studio | Try It Now | Content creators and podcasters |
| Stitch | Turn your ideas into UI with a vibe design tool | Try It Now | Anyone sketching an app or web screen |
| Opal | Build, edit and share AI mini-apps with natural language | Try It Now | Non-developers who want small tools |
| Jules | An asynchronous coding agent for bug fixes, writing tests and adding features | Try It Now | Teams with a developer |
| Stax | An AI evaluation toolkit that tests models and prompts | Try It Now | Anyone choosing between AI models |
| Playground | Describe a game idea and create a custom game in minutes, no coding required | Try it now | Marketers testing playful campaigns |
| Learn Your Way | Turns content into a learning experience tailored to you | Try It Now | Trainers and course creators |
| CC | An experimental group agent to help run a busy household | Try It Now | Families, not businesses |
| Dreambeans | Personalised collections of stories each day | Try It Now on the banner, Learn More on its tile | Personal reading |
| Google AI Edge Foresight | An on-device meeting companion for notes and recalling key details | Learn More | People in back-to-back meetings |
| Google AI Edge Eloquent | Turns raw, unpolished speech into well crafted text | Learn More | Anyone who drafts by talking |
| Disco | Tests AI features for the web, starting with GenTabs, which remixes open tabs into custom apps with Gemini 3 | Learn More | Heavy browser users |
| Literature Insights | Finds papers, structures data tables and creates reports and slide decks | Learn More | Researchers and analysts |
| Hypothesis Generation | A multi-agent research tool that simulates the scientific method | Learn More | Scientists |
| Computational Discovery | An agentic research engine that generates and scores code variations | Learn More | Research teams |
| Vantage | Develops and measures skills like collaboration and critical thinking through simulated teamwork | Learn More | Learning and development teams |
| Play with Putty | Collaborative vibe coding so teams can build tools and websites together in real time | Join Waitlist | Teams happy to wait |
Try It Now, Learn More, Join Waitlist: what each label tells you
The labels are the most honest signal on the page. Try It Now means you can open the experiment today. Learn More usually means you're reading about it rather than using it yet. Join Waitlist means you're queueing. The page also has filters across the top, All, Create, Develop, Explore and Learn, which help when you only care about making things or building things.
One small oddity worth knowing: Dreambeans shows Try It Now in the banner at the top of the page and Learn More on its own tile further down. Labels clearly change as experiments move along, so click through rather than trusting a screenshot someone shared last month. The Labs page also doesn't say what each experiment costs or where it's open, so check the experiment's own page before you plan around it.
Which experiments are worth a small business's time
For most business owners, three stand out. Stitch is a quick way to see a rough screen design for an idea before you brief a designer, and my Stitch walkthrough shows how it works. Opal lets non-developers build small AI mini-apps in plain English; before you build on it, read my Opal guide, which has its current status. Jules is the one to hand a developer, because it works through coding jobs in the background, and my guide to Jules covers the setup.
Stax deserves a mention too. It tests models and prompts, which is useful if you're arguing internally about which AI to use. Eloquent and Foresight sound handy for busy owners, but both are Learn More only, so they're not something you can lean on yet.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
How I'd test a Labs experiment without betting the business on it
- Pick one experiment labelled Try It Now that matches a job you already do every week.
- Use non-sensitive work only: a dummy product, an old blog post, a made-up customer. No client data.
- Give it one real task and judge the result against what you'd normally produce.
- Never build a client process on it. Experiments end, and Mixboard is proof that a Labs tool can shut down.
- Keep a copy of anything it makes outside the experiment, in your own files.
- If you want early warning of new tools, sign up for the Labs newsletter rather than checking the page every week.
My verdict: I'd try one Create or Develop experiment on non-sensitive work this month. Skip anything with a waitlist until it has its own help page, and skip anything you'd struggle to replace if it closed. Labs is the best free preview of where Google's AI is heading, as long as you treat it as a preview.
If you want a safe way to test new AI experiments without betting the business on them, that's the kind of AI enablement I help with. I run AI training and workshops for teams, and help businesses with AI enablement and workflow automation, so the busywork runs itself and people get time back. Most of that work is with SaaS marketing teams and small businesses. Here's how we can work together.
Frequently asked questions
What is Google Labs?
Google Labs is Google's home for AI experiments. It's a single page where you can try early AI tools, give feedback and sign up for early access to new experiments.
Which Google products started in Google Labs?
The Labs page lists Google Gemini (formerly Bard), Google Flow (formerly VideoFX), Gemini Notebook (formerly Project Tailwind), Google Flow Music (formerly ProducerAI) and Google AI Studio (formerly MakerSuite).
What does Join Waitlist mean on Google Labs?
It means you can't use the experiment yet and you're signing up to be let in later. Play with Putty carried that label as of Nov 2026.
Are Google Labs experiments safe for business use?
Treat them as previews. Use non-sensitive work, keep copies of anything they make, and don't build a client process on an experiment, because experiments can shut down.
How do I hear about new Google Labs experiments?
Sign up for the Labs newsletter on the Labs page, or apply to become a trusted tester to test what Labs builds next.
