Opera Introduces WebAI: Run AI Models in Your Browser Without the Cloud

Opera has introduced WebAI, a proposed web standard designed to let websites and browser extensions use artificial intelligence models directly on a user’s device instead of relying entirely on cloud-based services.

The feature is available as an early preview in Opera One, the company’s main browser. Opera says its approach supports different AI models, giving developers more flexibility when adding AI-powered features to websites and extensions.

How Does WebAI Work?

Traditional online AI services often send requests to remote servers for processing. With WebAI, compatible AI models can run locally on a computer, allowing certain tasks to be completed without sending the relevant information to an external AI server.

Opera says the technology can support several uses, including:

• Text generation: Create written content directly on the device.

• Content classification: Sort or identify different types of information.

• Information analysis: Identify relationships between pieces of content.

• Browser extensions: Add AI features to tools that organise tabs or analyse web pages.

The system is designed to work with multiple models rather than depending on a single AI provider.

Google and Apple Models Supported

Opera says WebAI can work with Google’s Gemma models. On Mac computers, it also supports Apple’s on-device foundation model.

The company plans to expand access to the feature across its browser products. Opera GX, which targets gamers, and Opera Neon, its AI-focused browser, are expected to receive similar support.

The current release is an early preview, and users must choose to enable the feature.

Could WebAI Improve Privacy?

Running AI models locally can reduce the need to send personal information to external servers. It may also allow users to access certain AI features without paying for a separate cloud AI subscription.

For developers, local processing could reduce the cost of running AI services because some requests would be handled by users’ own computers.

However, local AI does not automatically make every website or browser extension safe. Privacy also depends on how software accesses, stores and uses information. Users should still review permissions before allowing websites or extensions to process their data.

What Are the Limitations?

Local AI processing places more demands on a computer’s processor, memory and storage. The speed and quality of results can vary depending on the device and the model being used.

Some AI tasks may therefore run more slowly than they would on powerful cloud servers. Compatibility with different devices and models may also affect the experience.

WebAI is a proposed standard, so broader adoption will depend on how developers, browser makers and other technology companies support it.

What Happens Next?

Opera’s WebAI preview gives developers an opportunity to explore AI-powered websites and extensions that rely less on remote servers.

If support grows across browsers and websites, local AI could become a more common part of everyday browsing. For now, the technology remains in its early stages, and its wider impact will depend on performance, compatibility and adoption by developers.