Opera’s WebAI preview lets websites and extensions use compatible models on the user’s computer through browser interfaces. Activation is optional and the first release covers selected Windows and Mac hardware. The approach puts model execution inside the browser, giving developers tools for generating text, classifying information and comparing meaning without a separate local server.
Opera opened its WebAI preview in Opera One, Opera GX and Opera Neon. The feature lets websites and extensions call compatible AI models running on the user’s own computer. Opera places inference inside the browser, so developers need neither a separate AI application nor a configured local server or cloud-service key to use this path.[1], [2]
Users must enable WebAI in settings before a model downloads, and can switch it off afterward. Initial support covers Windows computers with Intel or AMD processors and Macs with Apple silicon. Linux, Intel Macs and Windows on Arm are outside the initial preview. Local processing does not require page content to be sent to a cloud AI provider; response speed still depends on the device’s computing resources.[1]
The Prompt, Decision and Semantic Embeddings interfaces support text generation, classification and comparisons of meaning. Opera prepared six experimental extensions, including tools that mark relevant article passages, make quizzes and visually group similar sentences. These are examples for developers rather than built-in browser functions. Opera describes WebAI as a proposed web standard; the preview does not establish a shared implementation across browsers.[1]
