ITHome October 28 news: browser maker Opera has introduced a new Web standard called WebAI, allowing users to run AI models directly in the browser without relying on cloud services.
Opera said WebAI is designed to be compatible with a variety of AI models. The company cited Google's Gemma as an example. This is an open-weights model launched by Google, first released two years ago. Opera also said that on Mac computers, WebAI supports calling Apple's on-device foundation models.
Opera emphasized that WebAI is not tied to any specific model, but is built for a variety of open-weights AI systems. The project grew out of Opera's earlier work on bringing locally running large language models into the browser.
Currently, Opera has launched this AI feature in its flagship browser Opera One as an early preview. The company said that Opera GX, the browser for gamers and streamers, as well as Opera Neon, a browser capable of using AI agents to perform tasks, will also receive the same feature in the near future. Users need to actively opt in to WebAI in order to use the related features.
Open-weights models offer users many benefits. First, such models can usually be downloaded and used for free. Second, their underlying model weights are made public, giving developers greater freedom to customize the models according to their own needs. However, unlike fully open-source software, open-weights models may still be subject to certain restrictions on use and modification.
Opera said privacy protection is also an important advantage of WebAI. Users can use AI features without sending personal information to corporate servers, and without paying for a cloud AI subscription.
Opera said in its announcement: "Developers can add AI features to their products without deploying AI infrastructure or paying for each request. Users do not need to buy an AI subscription service; information processed by local AI stays on users' own devices, and users retain control over the AI on their computers."
Opera also said that with WebAI, users can generate text directly in the browser, classify content, and understand the relationships between different pieces of information.
ITHome has noted that as AI assistants grow more capable and gain access to more and more personal information, privacy concerns are increasingly coming to the fore. When using AI chatbots, users often provide a great deal of personal information, sometimes without knowing how that data will be stored and used.
Locally running AI models offer an alternative, because all processing is done on the user's own device rather than being handed off to corporate cloud servers. This does not mean all privacy risks can be eliminated, but it does give users more control. For this reason, Opera has for years made digital privacy protection one of the main selling points of its products.
However, running AI models locally also involves certain trade-offs. Since the models need to directly use the computer's computing resources, they may increase the hardware load and thereby affect device performance. Opera cautioned that WebAI's actual performance will depend on the hardware configuration of the user's device.