Opera brings opt-in, on-device AI models to its browser

Opera is previewing WebAI, a standard for running AI models directly in its Opera One browser rather than sending requests to the cloud. Users can try text generation and content classification on their own device, while developers can add AI features without running model infrastructure or paying for each request.

Watch Desk analysis

What happened

Opera says WebAI is designed to support a range of AI models, including Google’s Gemma and Apple’s on-device foundation model on Macs. The company says the feature is opt-in and available as an early preview in Opera One. Opera GX and Opera Neon are expected to get it later. TechCrunch reports that WebAI grew out of Opera’s previous work on locally run browser models.

Our top picks

• Models run on the user’s device
  WebAI can generate text and classify content locally, without relying on a cloud model service.
• Developers can add AI without operating the infrastructure
  Opera says developers can build features without maintaining AI infrastructure or paying for every request.
• Choice of model is part of the pitch
  Opera says WebAI is not tied to one model, naming Gemma and, on Macs, Apple’s on-device foundation model as examples.
• The preview is opt-in
  Opera One users choose whether to enable it; Opera GX and Neon are due to follow.
• Your computer does the work
  Local processing avoids sending the information handled by the model to a cloud provider, but performance depends on the device’s hardware.

Why it matters

Browser-based local AI could give people a way to use model features without a cloud subscription or sending prompts to a company’s servers. It also gives developers another place to build AI tools, without paying for every request to a hosted model. The trade-off is practical rather than mysterious: the computer doing the work needs enough capacity, and its hardware affects performance.

Our read

This is a meaningful move towards making local AI a browser capability, not just something for people comfortable installing and configuring separate tools. The useful test is whether WebAI feels straightforward and performs well on ordinary devices. Privacy is a genuine advantage of local processing, though it is not a magic shield against every risk.

What to watch

• Which models Opera supports, and whether users can add their own.
• How performance varies across everyday hardware.
• When WebAI reaches Opera GX and Opera Neon.

Discussion spark: Would you choose slower, device-bound AI to keep processing local, or is cloud convenience worth sending your prompts to a provider?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.