Opera is going all-in with pushing AI for web browsing and it just might have delivered a solid knockout punch. The latest from the Norwegian company is Opera Neon, which puts an AI assistant at your disposal and also serves a handful of clever tricks to take actions across tabs so that you can avoid the boring and tedious parts of your workflow.
The big shift
Let’s start with the agentic capability called Do in Opera Neon, which is somewhat similar to the Operator agent introduced a few months ago. The overarching idea is to just say the task at hand, and the AI will autonomously handle it for you. It even has on-screen content awareness and lives as a sidebar window in the same tab to assist you.
For example, it can autonomously perform a deep web analysis and create a long report, somewhat like the Deep Research tool you get with Gemini and ChatGPT. Neon Do can also do a lot more, such as fill out forms on users’ behalf and pause at the crucial points where it needs user input. Users can also choose to take over control at any stage.
The sidebar is where you can essentially talk with the talk. For example, you can give a command like “pull items described from this DIY YouTube video and order them.” Neon Do will pull up all the information from the YouTube tab, extract the items described, and place an order. For low-stakes scenarios, it can help quickly discover information, pull details from a web page, and more — without having to open another tab.
The meaningful bits
Next, we have Tasks in Opera Neon. Think of them as a dedicated environment where multiple tabs are grouped and action is taken across. For example, you can have the AI analyze and create a table comparing products open across a dozen Amazon tabs. You can choose to take across different services (Gmail, Notion, etc.) that are open across tabs.
Next, we have Cards. Think of them as shortcuts with a predefined action the moment you invoke them. They are similar to how you interact with shortcuts in Perplexity’s Comet and Skills in Dia. In Opera’s case, they appear as beautifully designed cards, reminiscent of Pinboards in the main Opera browser.
Cards (or shortcuts) are my favorite part of AI web browsers as they speed up mundane tasks and save a heck of a lot of time. Creating them is pretty easy, as you just have to describe the task and it will be coded as an activity. You can give these cards unique names, and summon them with a simple “/“ command. There will also be a store where users can try out the cards created and shared by the community.