What is Agentic Browser?
At the beginning of this year, I wrote about OpenAI's Operator. A web browser with the combination of AI and Playwright to simulate human-like browsing, typing and clicking to complete a certain task on the web browser. This new type of web browser is called Agentic Browsers.
Today, I decided to do a survey on the current landscape of the Agentic Browser. This tool caught my eye:
BrowserOS
BrowserOS is an open-source Chromium-based browser. You get a browser with an AI chat panel. You can ask the AI to navigate to a website, find a specific button and click it to complete a task.
Today, many of our tasks are performed online using a web browser, such as making payments, replying to emails, and scheduling meetings. We can streamline these activities with an interactive browser that can open multiple tabs. For example, it can display your email inbox in one tab and your calendar in another. It can check your inbox to find someone who has agreed to your deal and wants to meet you online at a specific date and time. Meanwhile, in the second tab, it can schedule the appointment for you.
BrowserOS has an MCP server built in. That means your AI agent running in your terminal or other agentic application can access BrowserOS and automate tasks in the browser.
Another feature to highlight is workflow creation, which allows you to automate repetitive tasks to an AI agent in your browser.
Workflow creation and testing it.
Overall, BrowserOS will simplify your daily web-based tasks with automation and also as a developer to test the latest change directly in the browser.
Looking to automate some tasks?
Give it a try! https://www.browseros.com/
Disclaimer: I'm not affiliated with BrowserOS.
Member discussion