Opera has launched early access to WebAI, a technology enabling websites and extensions to run artificial intelligence directly on user devices. This approach eliminates the need for cloud-based requests, server infrastructure, or API keys, allowing developers to integrate AI features without incurring costs for query processing.
The feature is currently available in the Opera One browser, with plans to expand to Opera GX and the Neon browser. WebAI supports models such as Google's Gemma and native LLMs on Apple silicon. Users can enable the functionality via the browser settings, which triggers a model download process.
The system is compatible with 64-bit Windows computers using Intel or AMD processors, as well as Macs with Apple silicon. Opera notes that performance depends on local hardware specifications. The company emphasizes that this local processing model provides users with greater control over their data privacy.
WebAI is based on a proposed web standard currently under discussion with industry partners like Google and Microsoft. Microsoft has already integrated similar interfaces into Edge test versions. Opera previously experimented with over 2,000 local LLMs in 2024, concluding that advancements in open-source models now make local execution practical.
