Opera Launches WebAI, a New Web Standard for Local AI Processing in Browsers

Opera has introduced WebAI, a web standard that enables AI models to run inside Opera's browser without cloud processing. Opera is offering the features as an early preview in Opera One, Opera's flagship browser. Opera says users must opt in to access the features.

Opera says Opera GX and Opera Neon will receive the features soon. Opera GX targets gamers and streamers. Opera Neon uses AI agents to carry out tasks. WebAI builds on Opera's earlier efforts to integrate locally running models into Opera browsers.

Opera says WebAI can support many open-weight AI systems rather than depending on a particular model. Opera identifies Google's Gemma as one example. Google introduced Gemma, an open-weight model, two years ago. Opera says WebAI also supports Apple's on-device foundation model on Macs. Opera says WebAI enables users to produce text, categorize content and identify relationships between information within the browser.

Open-weight models are generally available to download and use for free. Public access to model weights gives developers greater freedom to customize model behavior. However, open-weight models retain some restrictions that distinguish the models from fully open-source software.

Opera says local AI lets users access features without transmitting personal information to company servers or buying cloud AI subscriptions. Opera says developers can introduce AI features without maintaining AI infrastructure or incurring charges for each request. Opera says users retain control over AI running on users' computers.

Privacy concerns are increasing as AI assistants become more capable and obtain access to personal information. Users frequently share extensive personal information with AI chatbots, sometimes without clearly understanding data storage or use. Local models process information on the user's device rather than a company's cloud server. Local processing does not remove every privacy risk, but local processing gives users somewhat greater control. Opera has emphasized digital privacy as a selling point for years.

However, local AI models increase demands on the user's computer and can affect performance. Opera says model performance depends on the device's hardware.