<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>MultiAgentOS blog and guides</title>
    <link>https://www.multiagentai.ai/blog/</link>
    <description>Articles and setup guides on local-first AI, multi-agent orchestration, BYO LLMs, Ollama, MCP tools, and desktop automation.</description>
    <language>en-us</language>
    <atom:link href="https://www.multiagentai.ai/rss.xml" rel="self" type="application/rss+xml" />
    <managingEditor>support@multiagentai.ai (MultiAgentAI)</managingEditor>
    <copyright>Copyright 2026 MultiAgentAI</copyright>

    <item>
      <title>Local AI agent on Windows: a practical 2026 setup guide</title>
      <link>https://www.multiagentai.ai/blog/local-ai-agent-windows-setup-2026/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/blog/local-ai-agent-windows-setup-2026/</guid>
      <pubDate>Mon, 25 May 2026 12:00:00 GMT</pubDate>
      <category>How-to</category>
      <description>Set up a Windows desktop AI agent with local models, files, screenshots, terminal commands, MCP tools, and supervised actions.</description>
    </item>

    <item>
      <title>Best local LLM models for desktop agents in 2026: how to choose</title>
      <link>https://www.multiagentai.ai/blog/best-local-llm-models-2026/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/blog/best-local-llm-models-2026/</guid>
      <pubDate>Mon, 25 May 2026 12:00:00 GMT</pubDate>
      <category>Local models</category>
      <description>Choose local LLMs for desktop agents by memory fit, context behavior, coding ability, tool discipline, latency, and cloud fallback.</description>
    </item>

    <item>
      <title>MCP vs function calling explained for desktop AI agents</title>
      <link>https://www.multiagentai.ai/blog/mcp-vs-function-calling-explained/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/blog/mcp-vs-function-calling-explained/</guid>
      <pubDate>Mon, 25 May 2026 12:00:00 GMT</pubDate>
      <category>Tools</category>
      <description>Understand MCP, function calling, and how desktop AI agents should scope tools safely.</description>
    </item>

    <item>
      <title>Set up Ollama for local AI agents</title>
      <link>https://www.multiagentai.ai/guides/ollama-local-ai-agents/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/guides/ollama-local-ai-agents/</guid>
      <pubDate>Tue, 19 May 2026 12:00:00 GMT</pubDate>
      <category>Guide</category>
      <description>Install Ollama, run a local model, verify the local server, and connect it to MultiAgentOS.</description>
    </item>

    <item>
      <title>Add an OpenAI API key to a desktop AI agent</title>
      <link>https://www.multiagentai.ai/guides/openai-api-key-desktop-agent/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/guides/openai-api-key-desktop-agent/</guid>
      <pubDate>Tue, 19 May 2026 12:00:00 GMT</pubDate>
      <category>Guide</category>
      <description>Create, store, test, and route an OpenAI API key safely in MultiAgentOS.</description>
    </item>

    <item>
      <title>Choose local GGUF models for desktop AI agents</title>
      <link>https://www.multiagentai.ai/guides/local-models-gguf/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/guides/local-models-gguf/</guid>
      <pubDate>Tue, 19 May 2026 12:00:00 GMT</pubDate>
      <category>Guide</category>
      <description>Understand quantization, memory, context length, and model testing before using a local model in an agent workflow.</description>
    </item>

    <item>
      <title>Connect MCP tools to desktop AI agents</title>
      <link>https://www.multiagentai.ai/guides/mcp-tools-desktop-ai/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/guides/mcp-tools-desktop-ai/</guid>
      <pubDate>Tue, 19 May 2026 12:00:00 GMT</pubDate>
      <category>Guide</category>
      <description>Configure MCP servers with least privilege, test them safely, and expose useful tools to MultiAgentOS.</description>
    </item>

    <item>
      <title>Set up desktop automation for AI agents</title>
      <link>https://www.multiagentai.ai/guides/desktop-automation-ai-agent/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/guides/desktop-automation-ai-agent/</guid>
      <pubDate>Tue, 19 May 2026 12:00:00 GMT</pubDate>
      <category>Guide</category>
      <description>Prepare visible sessions, permissions, stop controls, scoped actions, and safe tests for desktop-control AI.</description>
    </item>

    <item>
      <title>Multi-agent AI explained - what it actually is, and when it's useful</title>
      <link>https://www.multiagentai.ai/blog/multi-agent-ai-explained/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/blog/multi-agent-ai-explained/</guid>
      <pubDate>Sat, 09 May 2026 12:00:00 GMT</pubDate>
      <category>Concepts</category>
      <description>The honest version: a multi-agent system is multiple LLM calls coordinated by a controller. Here is what that looks like in practice, when it beats single-prompt approaches, and when it is overkill.</description>
    </item>

    <item>
      <title>Local LLMs vs. cloud APIs - the 5-year cost reality</title>
      <link>https://www.multiagentai.ai/blog/local-llm-vs-cloud-api-cost/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/blog/local-llm-vs-cloud-api-cost/</guid>
      <pubDate>Sat, 09 May 2026 12:00:00 GMT</pubDate>
      <category>Cost</category>
      <description>For many agentic workloads, local GPUs can reduce long-term API spend. This article walks through the math, break-even points, and exceptions.</description>
    </item>

    <item>
      <title>Cursor vs. MultiAgentOS for privacy-sensitive teams</title>
      <link>https://www.multiagentai.ai/blog/cursor-vs-multiagentos-privacy/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/blog/cursor-vs-multiagentos-privacy/</guid>
      <pubDate>Sat, 09 May 2026 12:00:00 GMT</pubDate>
      <category>vs</category>
      <description>Cursor is a great editor. MultiAgentOS is a local-first desktop AI workspace for teams that need a different privacy and workflow shape.</description>
    </item>

    <item>
      <title>How to run AI agents locally with Ollama - a practical 30-minute setup</title>
      <link>https://www.multiagentai.ai/blog/run-ai-agents-locally-with-ollama/</link>
      <guid isPermaLink="true">https://www.multiagentai.ai/blog/run-ai-agents-locally-with-ollama/</guid>
      <pubDate>Sat, 09 May 2026 12:00:00 GMT</pubDate>
      <category>How-to</category>
      <description>From a clean machine to a multi-agent setup: installing Ollama, picking a tool-capable model, wiring it into MultiAgentOS, and running the first task.</description>
    </item>
  </channel>
</rss>
