Opera brings local AI to its browser with an opt-in WebAI preview

Opera One users can now opt into an early preview that runs AI models inside the browser rather than relying on cloud processing. Called WebAI, the release is designed to work with multiple models. Opera's pitch is more control over the information AI handles, without requiring users to pay for a cloud AI subscription.

The introduction builds on Opera's previous work bringing locally run models into its browsers. WebAI is the company's new standard for that approach: the model does its processing on the user's own machine, rather than sending that work to a company's cloud server.

Several models, three kinds of browser work

Opera describes WebAI as model-agnostic, meaning it is not tied to one AI model. The company names Google's Gemma as a supported example and says Macs can also use Apple's on-device foundation model. Its stated aim is to support a wide range of open-weight systems.

Open-weight models make their underlying model weights publicly available, giving developers more freedom to customize how they work. They are typically free to download and use. That does not make them equivalent to fully open-source software: restrictions can still apply.

Opera says WebAI supports three kinds of work directly in the browser:

• Generating text.
• Classifying content.
• Identifying relationships between pieces of information.

Developers can add AI features without running AI infrastructure or paying for every request,

Opera, in its announcement quoted by TechCrunch

The computer takes on the cost

For users, Opera emphasizes where the information goes. The company says information handled by local AI stays on the device, and users remain in control of the AI running there. That offers an alternative to sending personal information to a server, though local processing does not eliminate every privacy risk.

The trade-off is a heavier workload for the computer. Running models locally puts more demand on the device and can affect performance. Opera says model performance depends on the user's hardware. The promise of avoiding cloud processing therefore comes with a practical constraint: the machine itself has to do the work.

For now, access is an opt-in early preview in Opera One, not an automatic addition for every user. Opera says its gaming-focused browser, Opera GX, and its AI-agent browser, Opera Neon, will receive the features later. Those versions remain planned additions rather than part of the current preview's availability.