Opera has launched WebAI, a new browser standard enabling users to run AI models directly on their machines without sending data to the cloud.
How it works
The standard is compatible with various open-weight systems. Opera cites Google’s Gemma, released two years ago, as a supported example. It also notes compatibility with Apple’s on-device foundation model on Macs.
The technology does not rely on a single provider. It builds on previous efforts to host local models within Opera’s interface. The feature is currently available as an early preview in Opera One. Opera GX and Opera Neon will receive the update shortly. Users must opt in to use the function.
Why use open-weight models
These models are typically free to download. Their weights are public, allowing developers to modify how they operate. They are not fully open-source software and carry some restrictions.
Privacy is the primary advantage. Users access AI capabilities without transmitting personal data to external servers or paying for cloud subscriptions.
“Developers can add AI features without running AI infrastructure or paying for every request,” the company said. “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.”
The system enables text generation, content classification, and relationship mapping of information entirely within the browser.
Privacy concerns
As AI assistants gain power and access to personal details, privacy issues have grown. Users frequently surrender vast amounts of data to chatbots without understanding how that information is stored or used.
Local processing offers an alternative. The computation occurs on the user’s hardware rather than a company’s server. This does not remove all privacy risks but provides greater control. Opera has long marketed digital privacy as a core strength.
Performance limits
Running models locally places higher demands on a computer. Performance depends on the device’s hardware. Users should expect trade-offs regarding speed and resource usage.
