Opera introduces WebAI standard for local browser AI

Opera has introduced WebAI, a web standard for running AI models directly in a browser without relying on the cloud. The company says the standard is designed to support a range of models and is not tied to any particular one.

Opera cites Google’s Gemma, an open-weight model Google first introduced two years ago, as one example. On Macs, WebAI also supports Apple’s on-device foundation model, according to Opera. The standard builds on the company’s earlier work to bring locally run models to its browsers.

The features are available as an early preview in Opera One. Opera says its gaming and streaming browser Opera GX, and its AI-agent browser Opera Neon, will get the features soon. Users must opt in.

Opera says WebAI can generate text, classify content and identify relationships between pieces of information in the browser. Running models on a user’s device can avoid sending the information they process to a company’s servers, and developers can add AI features without operating AI infrastructure or paying for each request, according to the company. Open-weight models make their weights publicly available, though they may still have restrictions and are not necessarily fully open-source.

Local processing can give users more control over where information is handled, but it does not remove every privacy risk. Running models locally also places demands on a computer, and Opera says performance depends on the device’s hardware.