<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Browser Agents on Andreas Nissen</title><link>https://andreasnissen.dev/tags/browser-agents/</link><description>Recent content in Browser Agents on Andreas Nissen</description><generator>Hugo</generator><language>en-us</language><copyright>© Andreas Nissen</copyright><lastBuildDate>Mon, 31 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://andreasnissen.dev/tags/browser-agents/index.xml" rel="self" type="application/rss+xml"/><item><title>Screen Use vs WebMCP: What Changes When an Agent Gets Governed Tools</title><link>https://andreasnissen.dev/writing/screen-use-vs-webmcp/</link><pubDate>Mon, 31 Aug 2026 00:00:00 +0000</pubDate><guid>https://andreasnissen.dev/writing/screen-use-vs-webmcp/</guid><description>&lt;p&gt;Screen use and WebMCP solve different parts of the agent-interface problem.&lt;/p&gt;&#10;&lt;p&gt;Screen use lets an agent interpret and operate the interface a person already uses. WebMCP lets the application expose typed operations designed for an agent to call. One starts from presentation. The other starts from an action contract.&lt;/p&gt;&#10;&lt;p&gt;That distinction changes what the system can validate, observe, and control. It does not make one approach universally better. It tells us where each approach belongs.&lt;/p&gt;</description></item></channel></rss>