<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>AI Tools on Vibhu Bhatnagar — PowerShell &amp; Infrastructure Engineer</title><link>https://pwsh.in/tags/ai-tools/</link><description>Recent content in AI Tools on Vibhu Bhatnagar — PowerShell &amp; Infrastructure Engineer</description><generator>Hugo -- gohugo.io</generator><language>en</language><lastBuildDate>Fri, 24 Apr 2026 23:36:10 +0530</lastBuildDate><atom:link href="https://pwsh.in/tags/ai-tools/index.xml" rel="self" type="application/rss+xml"/><item><title>Cut MCP Token Costs 88% with jmunch-mcp</title><link>https://pwsh.in/posts/cut-mcp-token-costs-jmunch/</link><pubDate>Fri, 24 Apr 2026 23:36:10 +0530</pubDate><guid>https://pwsh.in/posts/cut-mcp-token-costs-jmunch/</guid><description>&lt;h2 id="introduction">Introduction&lt;/h2>
&lt;p>If you run Claude Desktop, Claude Code, or any MCP-enabled client through the day, you are burning tokens you never asked for. Every MCP server call that returns a large payload — GitHub issues, Firecrawl scrapes, search results — dumps the entire response into context whether your agent needed three lines or three hundred thousand tokens worth of data.&lt;/p>
&lt;p>&lt;a href="https://github.com/jgravelle/jmunch-mcp" target="_blank" rel="noopener">jmunch-mcp&lt;/a> is a transparent MCP proxy that sits in front of your existing MCP servers and fixes this without changing anything about how you work.&lt;/p></description></item></channel></rss>