Summary
Opera launched WebAI on October 27, 2026. It lets websites and browser extensions use AI models and these models run on your own computer. The first early preview is in Opera One and Opera GX and Opera Neon will get it later. The main change is very simple. Some AI work now needs no data center and no paid AI plan and no trip to a company server.
What Happened
Opera is a web browser company from Oslo, Norway. On October 27, 2026, it said it has started to ship local AI models in its browsers. The first part of this push is called WebAI and it is now in early preview in Opera One, the main desktop browser of Opera. Opera GX is its browser for gamers and Opera Neon is its browser built around AI agents and both are next in line. Opera has not given a date for those two yet.
WebAI is a proposed web standard and Opera says companies like Google and Microsoft back it. A web standard is a shared set of rules and browsers and websites can follow these rules. In this case the rules let a website or an extension ask the browser to use a local AI model.
Opera says its version is model-agnostic and that means it is not tied to one AI company. It is built around open models. At launch you can use Google Gemma models and on Macs you can also use the on-device foundation model of Apple. Opera says more open models will come later and they will be tuned for different devices.
This is not the first step of Opera here. In 2024 it added local models to an experimental developer build of Opera One and Opera says it was the first major browser to do that. Now it brings that same work to everyday users.
How Local AI Is Different
Most AI tools today work in the cloud. You type a prompt and the website sends your text to a company server. The server runs the model and sends an answer back. You often pay for that work with a subscription and the company also sees the text you send.
Local AI flips that flow. The model runs on your own laptop and the text you feed into a local feature stays on your device. You do not need an AI subscription for that local work. A developer also does not need to run AI servers or pay for each request.
Opera gives a clear example and I like this example. An extension could sort your tabs. It could flag clickbait. It could spot job posts that fit you. And it could do all that with a model on your laptop. No account. No fee. No data sent to an AI service for that task.
With WebAI, a website or extension can do three main things in the browser. It can generate text. It can classify content, which means it can sort things into groups. And it can work out how pieces of information relate to each other.
What You Can Try in the Preview
WebAI is opt-in and that is an important detail. Nothing is downloaded until you turn it on. You also get site-specific permissions and that means you can decide which websites can access local AI on your machine. You stay in control of what runs and where it runs.
Opera has also released example extensions to show what developers can build. One turns an article into moving art, with similar sentences as visual neighbours. Another sorts your tabs. Another flags clickbait. Reports on the launch also name a quiz tool and a project tool. These are examples, not finished products for every user.
Opera also launched HTML-in-Canvas on the same day. It lets developers put live web pages inside 2D and 3D scenes and menus, buttons, and forms still work inside the scene. This part is mainly for developers. For most readers, WebAI is the news that matters most.
Which Devices Work at Launch
At launch, the WebAI preview works on 64-bit Windows PCs with Intel or AMD chips. It also works on Macs with Apple silicon. That covers most new Windows laptops and newer Macs.
Performance depends on your hardware. A fast, new laptop will handle local models better than an old, weak one. If your laptop is slow now, local AI may feel slow too. Test it on one small task first, before you rely on it for daily work.
Opera’s boss for browsers and AI, Krystian Kolondra, put the idea in one line: “A lot of what people want from AI doesn’t need a data center. Your laptop can already do it, without sending your data anywhere.”
Why It Matters for Freelancers
Freelancers juggle many tabs, clients, and files each day. A local AI tool that sorts tabs or flags low-value links can save small bits of time all day long. Those small bits add up. Less tab mess means less time lost when you switch back to client work.
Privacy matters too. Freelancers often work with client drafts, rates, and private notes. A local feature keeps that text on the device for that task, instead of sending it to a server. You still need to check what each extension does. But for local tasks, the privacy gain is real.
Cost matters as well. Many freelancers pay for several AI tools each month. A local feature needs no AI subscription for that task. It will not replace a strong cloud model for hard work. But for small, repeat tasks, it can cut costs.
If you use a browser a lot for client work, see how our guide to ChatGPT Chrome tabs compares with this new browser-first idea. The two tools take a different path to the same goal: less busy work in the browser.
Why It Matters for Students
Students often work on a tight budget. A free, local AI feature is easy to welcome. You do not need a paid plan to sort research, group notes, or turn an article into a quiz for revision. A quiz made from your own reading can help you test what you recall, instead of just reading the same page again.
Students also share laptops or work on campus Wi-Fi. A local tool does not send your study text to a server for that task.
But students need a fair warning too. Local AI is only as strong as the device and the model. A small, open model will not write a full essay as well as a large cloud model. Use it for sorting, grouping, and quick drafts. Check every fact yourself. For free tools that fit a student budget, see our list of free AI tools for students.
What to Watch Before You Switch
First, this is an early preview, not a finished release. Early previews can have bugs. Features can change. Do not move all your work to it on day one.
Second, websites must support WebAI for it to work on their pages. At launch, only some sites and extensions will use it. Do not expect every website you visit to offer local AI on day one.
Third, the model choice is small at launch. You get Gemma, and Apple’s on-device model on Macs. More models are planned, but they are not here yet. If you need a very strong model for hard writing or coding, a cloud tool may still serve you better.
Fourth, Opera GX and Opera Neon do not have WebAI yet. If you use one of those browsers, you will need to wait. Opera has said they are next, but it has not shared a date.
Fifth, check permissions for each site and extension. Opt-in is a good start. But each extension still needs your trust. Read what it asks to access. Turn off access for any site you do not trust.
How to Try It
Start small. If you use Opera One on a supported Windows PC or a Mac with Apple silicon, turn on the WebAI preview. Try one example extension first. Sort a messy set of tabs, or flag clickbait on a news page. Watch how fast it runs on your machine.
Then try one real task. Group research links for one project, or make a quiz from one article you need to revise. If it saves you time on that task, keep it for that task. If it feels slow, wait for the next update.
Full honesty from my side. I have not tested WebAI on my own laptop yet, so I cannot tell you its real speed from my own use. The idea looks strong on paper, and the real test is your machine and your daily work.
Opera’s bet is clear. A lot of everyday AI work does not need a huge data center. Your laptop may already be strong enough for it. The next few months will show if websites and developers agree, and if users pick local AI for the small jobs they do each day.