Browser Use launches Browser Use Box with persistent logins and Telegram control

• browser_use's launch post framed Browser Use Box, or bux, as a way to give browser agents a "24/7 computer" instead of tying runs to a local machine.
• In browser_use's product intro, the company said bux is a personal agent box powered by Browser Harness, with the pitch aimed squarely at long-running browser automation.
• WesRoth's demo summary said the hosted setup uses a real Chrome browser, keeps logins persistent, and keeps running after a laptop is closed.
• The same WesRoth walkthrough showed Telegram as the control surface, so commands can be sent to the browser agent from a phone while the remote session keeps working.

You can jump straight to browser_use's intro post, watch WesRoth's demo video, and compare that with the shorter launch line that reduces the whole pitch to one idea: browser agents finally get an always-on machine.

Browser Use is packaging browser automation as a hosted box instead of a local app. The core claim in browser_use's intro and its shorter launch post is continuity: the agent keeps a browser open and working around the clock.

That is a small product framing shift, but a useful one. The product is not pitched as a smarter model or a new agent loop. It is pitched as persistent runtime.

The launch language ties bux directly to Browser Harness. In browser_use's post, Browser Harness is the layer powering the box, while WesRoth's summary describes the hosted environment as a cloud machine running a real Chrome session.

The concrete mechanics surfaced so far are simple:

• hosted server environment, per WesRoth's summary
• real Chrome browser, per WesRoth's summary
• persistent logins across sessions, per WesRoth's summary
• ongoing execution even when the user's laptop is closed, per WesRoth's summary

That makes bux look less like another browser agent demo and more like remote infrastructure for browser-native tasks.

The most concrete interface detail in the evidence is Telegram. According to WesRoth's demo description, users can text commands to the agent from anywhere while the browser session stays live on the remote machine.

The video attached to WesRoth's post shows the split-screen workflow clearly: phone messages on one side, browser execution on the other. That gives Browser Use Box a very different feel from agent products that still assume the operator is sitting in front of the machine that owns the browser session.