<?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: Brandon Bennett</title>
    <description>The latest articles on DEV Community by Brandon Bennett (@agentsystems).</description>
    <link>https://dev.to/agentsystems</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.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3552200%2F5ef5db97-2960-4c93-9225-9e1f14ea81eb.jpg</url>
      <title>DEV Community: Brandon Bennett</title>
      <link>https://dev.to/agentsystems</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/agentsystems"/>
    <language>en</language>
    <item>
      <title>Federated app store for self-hosted AI agents (Apache-2.0)</title>
      <dc:creator>Brandon Bennett</dc:creator>
      <pubDate>Wed, 05 Nov 2025 06:31:30 +0000</pubDate>
      <link>https://dev.to/agentsystems/federated-app-store-for-self-hosted-ai-agents-apache-20-19ka</link>
      <guid>https://dev.to/agentsystems/federated-app-store-for-self-hosted-ai-agents-apache-20-19ka</guid>
      <description>&lt;p&gt;Self-hosted app store for AI agents. Federated discovery, container isolation, run on your infrastructure.&lt;/p&gt;

&lt;p&gt;The problem: most organizations either build every agent in-house or send their data to third-party servers.&lt;/p&gt;

&lt;p&gt;You wouldn't build your own email client or maps app - you'd download one. AI agents should work the same way. I spent a year building the infrastructure to make that possible: discover agents built by others, run them on your infrastructure (private cloud, on-premises, or local) without sending your data to third-party servers.&lt;/p&gt;

&lt;p&gt;Key architecture:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Federated Git-based index (fork-based ownership, no gatekeepers)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Container isolation + egress proxy (you configure which URLs agents can access)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Credential injection (API keys configured on host, not in agent images)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Model abstraction (works with Ollama local, cloud APIs, or hybrid)&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Hash-chained audit logs&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The platform works. The agent index is mostly empty, but someone has to build the rails so the trains can run.&lt;/p&gt;

&lt;p&gt;Apache-2.0 open source. Pre-release but functional.&lt;/p&gt;

&lt;p&gt;GitHub: &lt;a href="https://github.com/agentsystems/agentsystems" rel="noopener noreferrer"&gt;https://github.com/agentsystems/agentsystems&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Docs: &lt;a href="https://docs.agentsystems.ai" rel="noopener noreferrer"&gt;https://docs.agentsystems.ai&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Looking for:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;&lt;p&gt;Agent builders to publish to the index&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Security researchers to review the architecture&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Organizations that need self-hosted AI infrastructure&lt;/p&gt;&lt;/li&gt;
&lt;/ul&gt;

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