<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Open Web on Thibault Meunier</title><link>https://thibmeu.com/tags/open-web/</link><description>Recent content in Open Web on Thibault Meunier</description><generator>Hugo</generator><language>en</language><lastBuildDate>Thu, 27 Aug 2026 12:00:00 +0200</lastBuildDate><atom:link href="https://thibmeu.com/tags/open-web/index.xml" rel="self" type="application/rss+xml"/><item><title>Probing the MCP registry for anonymous access</title><link>https://thibmeu.com/posts/mcp-registry-analysis/</link><pubDate>Thu, 27 Aug 2026 12:00:00 +0200</pubDate><guid>https://thibmeu.com/posts/mcp-registry-analysis/</guid><description>&lt;p&gt;The Model Context Protocol, or MCP for short, is &lt;a href="https://modelcontextprotocol.io/docs/getting-started/intro" class="external-link" target="_blank" rel="noopener"&gt;&amp;ldquo;an open-source standard for connecting AI applications to external systems&amp;rdquo;&lt;/a&gt;.&lt;/p&gt;&#10;&lt;p&gt;In this post, I briefly try to understand how open such an ecosystem is. I&amp;rsquo;ll use a very simple definition: as a client, does a server require me to log in, does it have a login facility, or is it completely open?&lt;/p&gt;&#10;&lt;p&gt;The result is that authentication is common in the part of the MCP ecosystem I measured. About 28% of remote endpoints required it to list tools, while another 23% advertised it. I do not have a good comparison with the Web, even though this seems quite high.&lt;/p&gt;</description></item></channel></rss>