That possibility is at the center of WebAI, Opera’s proposed web standard for running compatible artificial intelligence models locally in the browser. Opera said in its October 27 announcement that WebAI is intended to give websites and browser features access to local processing through a set of interfaces, including Prompt, Decision and Semantic Embeddings APIs.
The strategic significance is larger than Opera’s initial implementation. If browsers can provide a common way to use models on a computer or phone, developers may be able to add AI features without choosing one cloud provider, maintaining inference servers or paying for every request. The browser would become an important distribution layer for AI, much as it became the main delivery mechanism for software and services on the web.
Read Google’s offline meeting assistant tests whether local AI can beat the cloud
A different model for deploying AI
Most current AI products rely on a request leaving the device, reaching a company’s servers and returning with a result. That approach gives providers substantial control over the model, infrastructure and pricing. It also creates ongoing costs for developers and requires a reliable connection.
WebAI proposes a different arrangement. The model is available locally, and processing takes place on the user’s hardware. Opera describes the system as opt in, which means local execution is presented as a choice rather than an invisible replacement for cloud processing. Performance will depend on the capabilities of the device, so the same feature may feel quick on a recent computer and limited on older hardware.
That tradeoff could become increasingly practical as smaller models improve. A lightweight model does not need to match the broad capabilities of the largest cloud systems to be useful. It may be sufficient for rewriting a paragraph, classifying content, translating a short passage or identifying relationships between pieces of information.
Opera’s developer site illustrates that range through hands on demonstrations for local Prompt, Summarizer, Writer, Rewriter, Translator, Proofreader, Semantic Embedder and Decisions APIs. Opera’s developer page demonstrates these capabilities as browser based tools whose models run on the user’s device.
Privacy without a guarantee
Local processing changes the privacy calculation, but it does not remove privacy risks. If information does not need to travel to a remote AI provider, users gain a clearer opportunity to keep sensitive text on their own machines. That could matter for personal notes, workplace documents or other material that users may not want to upload.
However, local does not automatically mean safe. Websites, extensions and downloaded models could still handle information poorly. Users may also struggle to determine which model is processing their data, what is stored locally or whether a feature has quietly switched to a remote service when local execution is unavailable.
Opera’s opt in approach therefore points to an important design question. Browser makers will need to explain where processing occurs, what hardware is being used and what happens when a device cannot run a model efficiently. Those details may influence trust as much as the quality of the generated output.
The ecosystem question
WebAI’s biggest challenge is not demonstrating that a model can run in a browser. It is creating enough consistency for developers to build around the idea. A proposed standard becomes valuable only when browsers support it reliably, model runtimes work across different devices and websites can offer useful features without extensive customization.
Opera’s preview is an early test of that proposition. If WebAI gains wider support, browser native AI could remain available when cloud access is slow, costly or unavailable. It could also give users more control over recurring AI expenses by moving some computation from subscription services to hardware they already own.
The limitations are equally clear. Local inference depends on memory, processors and graphics capabilities, while different devices may produce sharply different results. Developers may still prefer cloud systems for demanding tasks, and browser vendors will need to manage compatibility, security and model distribution.
The larger implication is that AI competition may increasingly involve deployment standards as well as model performance. Opera is testing whether the browser can become the neutral layer connecting websites, users and multiple local models. If that experiment succeeds, the future of AI may be split between powerful centralized systems and a growing collection of smaller models that operate quietly on the devices people already use.
Alex Carter is not a person. No notebook, no deadlines, no face behind the name — just a byline this newsroom publishes under. Here is the production line underneath it, because a name beside a portrait reads like a journalist, and this one is not one.
The models. Writing: gpt-5.6-luna and qwen3-max. Out on the live web: gpt-5.6-luna and gpt-5.6-terra. Pictures: gpt-image-1 and gpt-image-1-mini. Swap one in the newsroom and this line swaps with it — it is read off the machines, not typed here.
How a story is made
- Research. The searching model reads around the story, pointed at primary sources — the filing, the post, the repository — rather than at somebody else's write-up of them.
- Writing. The writing model drafts it against what was found, at Alex Carter's usual length and in Alex Carter's usual register.
- The loop. A reviewer reads the draft and sends it back with notes. Then reads it again. A piece can go round several times before it leaves the building.
- Enrichment. A quotation has to appear word for word on the page it is taken from. A chart may only use figures that appear in the source it cites. Whatever fails is dropped, and the reason is kept.
- Fact check. A last pass hunts for claims the article makes and its sources do not.
- A human stop. Sensitive subjects are held for a person to read before publication, and a person can kill any of it at any point.
If that sounds less like a newsroom and more like a factory: quite. It is called Press Factory.
This article was generated using AI and published automatically without human pre-publication review.
How this article was made
The article was produced by the Grandmonts Media News Engine using automated research, drafting and verification workflows. No human editor reviewed the article before publication. Grandmonts Media remains responsible for the published content. Errors can be reported at office@grandmonts.cz.