Opera’s October 27, 2026 WebAI announcement includes Project Lens, an experimental extension that colors article paragraphs according to their relevance to a project brief. It offers a concrete way to try local browser AI. Before using the colors to skip reading, test whether they preserve the details your decision needs.
Opera’s WebAI announcement and example extensions
Check the preview and extension separately
Opera’s published trial instructions start with Opera One: update the browser, then enable Web AI API under Settings → Opera AI. This downloads a compatible local model. The preview supports 64-bit Intel/AMD Windows PCs and Apple-silicon Macs; Opera excludes Intel Macs, Linux and Windows on Arm. Its announcements differ on GX and Neon rollout, so this guide confines the trial path to Opera One.
The Project Lens listing identifies Opera Software as the developer and describes the extension as a demonstration requiring built-in AI to be enabled. The retrieved listing shows version 1.3.0, updated October 28. Record the version actually installed rather than assuming a later build behaves identically.
Project Lens listing, developer and version
Write the answer you need before trying the tool
For an original evaluation exercise, choose a public article whose full text you can inspect. Write a narrow brief, such as: “Find details that determine whether this proposed service can be used for a small pilot next month.” Before turning on highlighting, read the article and make your own answer sheet: availability date, eligibility, cost conditions and unresolved requirements.
Keep the first trial small enough to review manually. A short, clearly written public announcement is easier to assess than a large set of unfamiliar documents. Save the URL, retrieval date, exact brief and paragraph locations. Use public material for this exercise; it does not require private client records.
Look for relevant caveats that receive little attention
Imagine a fictional announcement with three paragraphs. Paragraph A promises a useful pilot. Paragraph B says access starts after the month in your brief and requires an invitation. Paragraph C describes a colorful launch event. For the stated task, the restriction in B is essential even if the launch promise in A sounds more exciting. C might be expendable.
Compare the extension’s treatment with your answer sheet. Record two kinds of mismatch: important material that is easy to overlook, and material emphasized despite having little bearing on the brief. This is a proposed exercise, not a reported result from Project Lens. Do not infer an error rate without a defined sample and actual observations.
Change one condition and compare again
Repeat with a different brief, such as preparing a launch-event summary. In the fictional example, paragraph C now matters. Keeping the document fixed while changing only the task helps you examine whether the highlights respond usefully to your purpose. Record the exact wording and observed result for each attempt.
Treat a highlighted passage as a reading aid. Relevance and truth are separate questions: a paragraph can be central to your project while making an unsupported claim. Open the original evidence before relying on dates, figures or promises, and read surrounding caveats before discarding an unhighlighted passage.
Opera describes the model processing as local. This article has not installed the extension, measured performance or audited its network behavior. A local-processing description does not substitute for a review of the complete workflow. Readers in the USA and India can use the same small evaluation method while checking their own browser build and device.
Judge the reading exercise
