Opera adds direct neural network launch in browser
The WebAI feature allows processing data on the user's computer without accessing cloud-based AI
Opera has begun implementing WebAI—a technology that enables websites and extensions to access neural networks running directly on the user's device. The first platform for the preliminary launch is the Opera One browser. The company plans to add these capabilities to Opera GX and Opera Neon in the future.
WebAI enables the creation of tools for text generation, content classification, and identifying semantic connections between materials. For example, an extension could filter suitable job vacancies, highlight publications of interest to the user, or assist in searching through saved materials. Developers do not need to set up their own cloud service to handle each request.
Opera's implementation supports various compatible models. Among the available options, the company lists Google's Gemma, and on Mac computers, the built-in Apple model. No separate application for local AI needs to be installed.
The feature is opt-in: the browser does not load models until activated. Users can also specify which websites are allowed to access the local AI. According to Opera, data processed by such models remains on the device.
The preliminary version is available on computers with 64-bit Windows and Intel or AMD processors, as well as on Macs with Apple Silicon chips. The speed of neural network operations depends on the computer's specifications.
WebAI itself currently has the status of a proposed web standard. Therefore, its appearance in Opera does not yet mean that such capabilities are available in all browsers.
