Opera has unveiled a new web standard called WebAI that enables artificial intelligence models to run directly inside its browser, eliminating the need for cloud-based processing. The announcement marks a significant step in the browser maker's effort to bring on-device AI capabilities to everyday users without requiring external server infrastructure or paid subscriptions.

The WebAI standard is designed with flexibility in mind, supporting a broad range of open-weight AI models rather than locking users into a single ecosystem. Opera cited Google's Gemma - an open-weight model that Google introduced roughly two years ago - as one compatible example. For users on Apple hardware, WebAI also supports Apple's on-device foundation model, giving Mac users an additional option for local AI processing.

Despite these specific examples, Opera has been clear that WebAI is not built around any one model. The standard is intended to accommodate a wide variety of open-weight AI systems, and it builds on earlier efforts by the company to integrate locally running models into its browser products.

Early Access Across Opera's Browser Lineup

Opera is currently rolling out the new AI capabilities as an early preview within its primary product, Opera One. The company has confirmed that two of its other browsers - Opera GX, which is tailored for gamers and streamers, and Opera Neon, which relies on AI agents to carry out tasks on behalf of users - will receive the same functionality in the near future. Participation in the early preview is opt-in, meaning users must actively choose to enable the feature rather than having it activated by default.

The Case for Open-Weight Models

Open-weight AI models carry several practical advantages for end users. They are generally free to download and use, which removes the cost barrier associated with many commercial AI services. Their model weights are made publicly available, giving developers meaningful latitude to modify and adapt them for specific use cases. That said, open-weight models are not entirely without restrictions - unlike fully open-source software, they typically come with certain usage limitations set by their creators.

One of the most prominent benefits Opera highlights is privacy. Because the AI processing occurs locally on the user's device, personal information does not need to be transmitted to a company's remote servers. Users also avoid the recurring costs of cloud-based AI subscriptions.

"Developers can add AI features without running AI infrastructure or paying for every request. Users don't need AI subscriptions, the information the local AI handles stays on their own device, and they stay in control of the AI on their computer."

Beyond privacy, Opera says WebAI supports a range of practical functions that can be performed entirely within the browser. These include generating text, classifying content, and identifying relationships between different pieces of information - all without leaving the browser environment or pinging an external server.

Privacy Concerns in the Age of AI Assistants

The push toward local AI processing comes at a time when privacy concerns surrounding AI tools have grown considerably. As AI assistants become more capable and gain deeper access to personal data, questions about how that information is stored, shared, and used have become increasingly prominent. Users frequently provide large volumes of personal information to AI chatbots, often without a full understanding of where that data ends up or how long it is retained.

Locally run models present a meaningful alternative by keeping all processing on the user's own hardware rather than routing it through a third-party cloud environment. While this approach does not eliminate all potential privacy risks, it does shift a degree of control back to the individual user. This aligns with a broader positioning strategy that Opera has pursued over the years, having long emphasized digital privacy as a core feature of its browser products.

Hardware Demands and Performance Trade-offs

Running AI models locally is not without its drawbacks. Executing these workloads directly on a user's device places greater demand on the machine's processor, memory, and - where applicable - its graphics hardware. This can have a noticeable impact on overall system performance, particularly on older or less powerful devices. Opera has acknowledged this limitation, noting that the quality and speed of model performance will vary depending on the specifications of the user's hardware.

The trade-off between privacy and performance is one that users will need to weigh for themselves. For those with capable hardware who prioritize keeping their data local, WebAI offers a compelling option. For others working on less powerful machines, the experience may be more constrained. As device capabilities continue to improve and AI models become more efficient, however, the gap between cloud-based and on-device AI performance is expected to narrow over time.