<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Bot2Bot Labs</title>
    <description>The latest articles on DEV Community by Bot2Bot Labs (@bot2botlabs).</description>
    <link>https://dev.to/bot2botlabs</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4235766%2F54fa1724-c074-441c-be0c-400c5ffba6be.png</url>
      <title>DEV Community: Bot2Bot Labs</title>
      <link>https://dev.to/bot2botlabs</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/bot2botlabs"/>
    <language>en</language>
    <item>
      <title>Daily snapshots of the MCP Registry: first numbers</title>
      <dc:creator>Bot2Bot Labs</dc:creator>
      <pubDate>Mon, 09 Nov 2026 19:19:07 +0000</pubDate>
      <link>https://dev.to/bot2botlabs/daily-snapshots-of-the-mcp-registry-first-numbers-5g3</link>
      <guid>https://dev.to/bot2botlabs/daily-snapshots-of-the-mcp-registry-first-numbers-5g3</guid>
      <description>&lt;p&gt;The official &lt;a href="https://registry.modelcontextprotocol.io" rel="noopener noreferrer"&gt;MCP Registry&lt;/a&gt; shows what’s listed &lt;em&gt;now&lt;/em&gt;, but it doesn’t keep a history. Since &lt;strong&gt;6 November 2026&lt;/strong&gt; I’ve been taking a full snapshot every day, plus read-only &lt;code&gt;tools/list&lt;/code&gt; probes of public hosted servers (initialize + list tools only — never calling a tool). This is a baseline from the &lt;strong&gt;7 November 2026, 06:36 UTC&lt;/strong&gt; snapshot (&lt;a href="https://github.com/bot2botlabs/mcp-shelf-archive/releases/tag/snapshot-2026-11-07" rel="noopener noreferrer"&gt;mcp-shelf-archive &lt;code&gt;snapshot-2026-11-07&lt;/code&gt;&lt;/a&gt;). Two days of history so far, so nothing here is a trend yet.&lt;/p&gt;

&lt;h3&gt;
  
  
  What’s on the shelf
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;57,993&lt;/strong&gt; entries: &lt;strong&gt;57,090&lt;/strong&gt; active and &lt;strong&gt;903&lt;/strong&gt; deprecated.&lt;/li&gt;
&lt;li&gt;Overnight from 6→7 November: &lt;strong&gt;158&lt;/strong&gt; added, &lt;strong&gt;3&lt;/strong&gt; removed.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;19,109&lt;/strong&gt; active entries have a latest-version date in October 2026; &lt;strong&gt;9,295&lt;/strong&gt; more are dated 1–7 November (that’s the date of each entry’s latest version, not when it was first listed).&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Who publishes
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;36,027&lt;/strong&gt; publishers.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;91%&lt;/strong&gt; list exactly one server.&lt;/li&gt;
&lt;li&gt;Ten publishers account for about &lt;strong&gt;15%&lt;/strong&gt; of active entries; the single largest GitHub-namespaced account lists &lt;strong&gt;2,366&lt;/strong&gt;.&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Source links and downloads
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;34.7%&lt;/strong&gt; of active entries give no source repository.&lt;/li&gt;
&lt;li&gt;Of &lt;strong&gt;858&lt;/strong&gt; linked GitHub repositories checked, &lt;strong&gt;104 (12%)&lt;/strong&gt; no longer exist; &lt;strong&gt;55%&lt;/strong&gt; of the rest have zero stars.&lt;/li&gt;
&lt;li&gt;Across &lt;strong&gt;4,643&lt;/strong&gt; servers whose npm packages we checked: median &lt;strong&gt;81&lt;/strong&gt; downloads/week; the top &lt;strong&gt;1%&lt;/strong&gt; take about &lt;strong&gt;90%&lt;/strong&gt; of all downloads in that sample.&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Tool lists (hosted, public HTTPS only)
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;Asked &lt;strong&gt;550&lt;/strong&gt; hosted servers; &lt;strong&gt;388&lt;/strong&gt; answered openly; &lt;strong&gt;123&lt;/strong&gt; required a login.&lt;/li&gt;
&lt;li&gt;Median server exposes &lt;strong&gt;9&lt;/strong&gt; tools; the largest in the sample, &lt;strong&gt;165&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Of &lt;strong&gt;6,621&lt;/strong&gt; tools listed, &lt;strong&gt;31%&lt;/strong&gt; carry no annotations at all (no readOnly / destructive hints).&lt;/li&gt;
&lt;li&gt;Of &lt;strong&gt;152&lt;/strong&gt; servers already seen the day before, &lt;strong&gt;16&lt;/strong&gt; changed tool definitions overnight — &lt;strong&gt;14&lt;/strong&gt; of those 16 were one publisher updating the same shared tools across its servers. Read that as one night’s data, not a rate.&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Method (short)
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;Registry: public MCP Registry API, latest version of every entry, daily.&lt;/li&gt;
&lt;li&gt;Tool lists: public, no-login, HTTPS endpoints; rotating sample + fixed watch list; &lt;code&gt;initialize&lt;/code&gt; and &lt;code&gt;tools/list&lt;/code&gt; only.&lt;/li&gt;
&lt;li&gt;Downloads / repos: official npm and GitHub APIs, daily rotating samples.&lt;/li&gt;
&lt;li&gt;No logins, no scraping, no tool calls.&lt;/li&gt;
&lt;/ul&gt;

&lt;h3&gt;
  
  
  Limits
&lt;/h3&gt;

&lt;p&gt;Samples rotate and aren’t the whole shelf. Counting entries isn’t judging quality. Corrections welcome: &lt;a href="mailto:jordan@bot2botlabs.com"&gt;jordan@bot2botlabs.com&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Full write-up with the same figures and limits:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://bot2botlabs.com/r/mcp-gh-202611-a" rel="noopener noreferrer"&gt;https://bot2botlabs.com/r/mcp-gh-202611-a&lt;/a&gt;&lt;/p&gt;

</description>
      <category>mcp</category>
      <category>ai</category>
      <category>opensource</category>
      <category>data</category>
    </item>
  </channel>
</rss>
