What is Opera WebAI? Most AI tools you use today work the same way. You type something in, it travels to a company's servers, a mo...
What is Opera WebAI?
Most AI tools you use today work the same way. You type something in, it travels to a company's servers, a model processes it, and the answer comes back. Opera is trying a different route. With WebAI, the model runs on your own computer, inside the browser, and your data never has to leave the machine.
WebAI is a proposed web standard that lets websites and browser extensions tap into AI models hosted by the browser itself. Opera says the proposal has backing from companies such as Google and Microsoft. The company announced it on October 27, 2026, as an opt-in early preview in Opera One, its flagship browser.
The idea is simple. If your laptop can already handle many everyday AI jobs, there is little reason to send your personal information to a data center and pay a monthly fee for the privilege.
Which AI models does WebAI work with?
Opera built WebAI to be model-agnostic. It is not tied to one provider and is designed around open-weight models, so users are not locked into a single AI company.
At launch, Opera names Google's Gemma as one example. Gemma is an open-weight model Google introduced about two years ago. On Macs, WebAI also supports Apple's on-device foundation model, which means Mac users can run Apple's own model directly inside Opera One.
The feature also works with compatible models already sitting on a user's device. Opera says it plans to add more open models tuned for different kinds of hardware over time.
How the idea grew out of Opera's earlier work
This did not appear out of nowhere. In April 2024, Opera added experimental support for local large language models to the developer stream of Opera One. That release covered roughly 150 model variants from about 50 families, including Meta's Llama, Vicuna, Google's Gemma, and Mistral's Mixtral. Each one typically needed between 2 and 10 GB of local storage, and they sat alongside Aria, Opera's online browser assistant.
Back then, local AI was a tool for developers willing to experiment. WebAI is the step that brings the same approach to regular users and gives web developers a proper way to build on it.
What can WebAI actually do?
According to Opera, WebAI lets anyone generate text, classify content, and work out how pieces of information relate to each other, all inside the browser. For developers, that translates into a set of building blocks:
Prompt API for generating text
Decision API for classifying information
Semantic Embeddings API for finding relationships between pieces of content Summarizer, Writer, Rewriter, Translator and Proofreader APIs, with playgrounds that run entirely on the device
Local inference is built into the browser. That means there is no separate AI app to install, no local server to set up, and no API key to hunt down.
Opera also released six experimental extensions to show what is possible. They are idea starters, not official Opera features:
Project Lens colors the paragraphs of an article by how much each one matters to a project you describe.
Got It? writes a short quiz about an article you just read.
Questionable Intelligence asks playful questions about a page and charts the verdict.
Semantic Art turns an article into moving art, with similar sentences drifting next to each other.
RecordIn Alchemy rewrites any article as an over-the-top inspirational social media post.
C·I·A·O has two ASCII cats improvise a scene on any topic.
Opera also pitched a more practical case: an extension that sorts your tabs, flags clickbait, or spots relevant job posts using a model on your laptop, with no account, no fee, and no data sent to an AI service.
Why local AI matters for privacy
Privacy is the strongest argument for running AI on your own hardware. As AI assistants get more capable, they also get more access to personal information. People hand over a surprising amount to chatbots, often without a clear picture of how that data is stored or used.
Local models change the equation because the processing happens on your device instead of a company's cloud server. That does not erase every privacy risk, but it gives you more control. Opera describes the benefits in plain terms: developers can add AI features without running AI infrastructure or paying for every request, users do not need AI subscriptions, and the information the local AI handles stays on their own computer.
It also fits Opera's long-standing brand. The company has treated digital privacy as one of its main selling points for years, so a push toward on-device AI is a natural extension.
What are open-weight models, and why do they help?
The phrase comes up constantly with WebAI, so it is worth a quick explanation. Open-weight models publish their underlying weights, the numerical settings that make the model work. They are typically free to download and use, and developers get more freedom to customize how they behave.
They are not the same as fully open-source software, though. Open-weight models still come with some restrictions on how they can be used, so the label describes access to the weights and does not mean anything goes.
You stay in control
Opera made WebAI opt-in. Nothing is downloaded until you turn the feature on in Settings, and you can switch it off at any time. Site-specific permissions let you decide which websites are allowed to use local AI, so a page cannot quietly start using your computer's resources.
Once you enable it, the browser downloads a compatible local model. Opera says that step may take a few minutes.
Availability: who can try it now?
WebAI is available as an early preview in Opera One. Opera says Opera GX, its browser for gamers and streamers, and Opera Neon, its browser built around AI agents that perform tasks, will get the same treatment soon.
Hardware support is limited for now. WebAI works on:
64-bit Windows PCs with Intel or AMD processors
Macs with Apple silicon
It is not available on Intel-based Macs, Linux, or Windows on Arm.
The trade-offs of running AI locally
Local AI is not free of costs. Running models on your own machine puts more demand on your computer, which can affect performance. Opera is upfront that results depend on your hardware. A recent laptop with plenty of memory will handle it far better than an older machine.
Opera says it will keep working to make local AI lighter and faster, and it calls this launch the start of a broader push. Expect more models and more developer tools to follow.
Local processing also has limits as a privacy fix. It reduces what leaves your device, but you still need to trust the extensions and sites you give permission to.
A side note: HTML-in-Canvas
Opera announced a second feature on the same day, aimed at gamers. HTML-in-Canvas lets real, interactive web pages sit inside 2D and 3D scenes. Previously, canvas content was just pixels you could not click or type into. With the new capability, menus, buttons and forms can move with a scene and still work. Opera sees it as useful for game interfaces such as shops, inventories, settings and chat, and it is most relevant to Opera GX.
Why WebAI could matter beyond Opera
If WebAI becomes a real standard, it could change how web developers think about AI features. Today, adding an AI tool to a website usually means paying for cloud requests, managing infrastructure, and handling users' data responsibly. A browser-level standard would let developers call a local model the same way they call any other browser feature.
For users, the pitch is just as clear. AI features without another subscription, without handing over personal details, and with a permission prompt that puts the decision in your hands.
Whether it catches on depends on how many browsers and developers adopt it, and on whether local models are good enough on everyday hardware. For now, Opera is the first to ship it, and anyone curious can try it in the early preview.
Frequently asked questions
What is WebAI?
WebAI is a proposed web standard that lets websites and browser extensions use AI models hosted by the browser and running on the user's own device instead of in the cloud.
Which browser supports WebAI?
It is available as an opt-in early preview in Opera One. Opera GX and Opera Neon are expected to get it soon.
Which AI models does WebAI support?
It is built to work with a wide range of open-weight models. Opera names Google's Gemma as one example, and Mac users can also run Apple's on-device foundation model.
Do I need an AI subscription to use it?
No. Because the model runs on your computer, there is no cloud subscription involved.
Is my data sent to the cloud?
Opera says the information the local AI handles stays on your own device. Local processing lowers privacy risk but does not remove every risk.
Is WebAI turned on by default?
No. You must enable it in Settings, nothing is downloaded until you do, and you can turn it off at any time.
Can I control which websites use it?
Yes. Site-specific permissions let you decide which websites can access local AI.
What computers does WebAI run on?
64-bit Windows PCs with Intel or AMD processors, and Macs with Apple silicon. Intel-based Macs, Linux, and Windows on Arm are not supported yet.
Will it slow down my computer?
It can. Local models demand more from your hardware, and performance depends on your device. Opera says it is working to make local AI lighter and faster.
What is the difference between open-weight and open-source models?
Open-weight models publish their weights and are usually free to download and customize, but they still carry some usage restrictions. Fully open-source software has fewer limits.
References
Unite - https://www.unite.ai/opera-launches-webai-bringing-local-ai-models-to-its-browsers/
Tech crunch - https://techcrunch.com/2026/10/27/opera-introduces-a-new-web-standard-to-let-users-deploy-ai-locally-in-their-browser/
Ground News - https://ground.news/article/opera-introduces-a-new-web-standard-to-let-users-deploy-ai-locally-in-their-browser
