<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Ai on Alex Lukens</title><link>https://alukens.com/tags/ai/</link><description>Recent content in Ai on Alex Lukens</description><generator>Hugo</generator><language>en-us</language><copyright>© alukens.com</copyright><lastBuildDate>Fri, 20 Nov 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://alukens.com/tags/ai/index.xml" rel="self" type="application/rss+xml"/><item><title>local-ai-search-provider: A GNOME Shell Extension for querying AI</title><link>https://alukens.com/posts/local-ai-search-provider/</link><pubDate>Fri, 20 Nov 2026 00:00:00 +0000</pubDate><guid>https://alukens.com/posts/local-ai-search-provider/</guid><description>&lt;div class='callout callout-custom' style='background-color: transparent; border: 3px solid #d340e0;'&gt;
 &lt;div class="callout-inner"&gt;
 
 ⚡️ &lt;u&gt;AI Disclosure&lt;/u&gt;
 &lt;br /&gt;
 This project was implemented fully with Generative AI, with human input for feature proposals &amp;#43; development. Notably, the Deepseek V4.1 Flash 1017 model, served from Opencode Go
 
 &lt;/div&gt;
&lt;/div&gt;

&lt;p&gt;&lt;a href="https://github.com/alexdlukens/local-ai-search-provider"&gt;local-ai-search-provider&lt;/a&gt; is a GNOME extension that adds integration for querying AI from the GNOME shell search (Super-key menu)&lt;/p&gt;
&lt;h2 id="overview"&gt;Overview&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Allows configuration of an OpenAI compatible endpoint (local or cloud). Optional API key authentication&lt;/li&gt;
&lt;li&gt;Configure model + system prompt&lt;/li&gt;
&lt;li&gt;Small set of tools that can be used for searching the web&lt;/li&gt;
&lt;li&gt;Streams responses inline in the GNOME Shell search interface, or in a dedicated &lt;code&gt;AI Chats&lt;/code&gt; application&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;In my case, I wanted a GNOME Shell extension that I could use with Unsloth Studio&amp;rsquo;s local API endpoint. I used the &lt;a href="https://huggingface.co/unsloth/gemma-4-E2B-it-qat-GGUF"&gt;Gemma4 E2B QAT&lt;/a&gt; model for the fast model load, prompt processing, and decode performance&lt;/p&gt;</description></item></channel></rss>