<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/ecf0342b2d83423a9790e84e82caacfd&quot; frameborder=&quot;0&quot; width=&quot;1920&quot; height=&quot;1440&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1440</height><width>1920</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1440</thumbnail_height><thumbnail_width>1920</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/ecf0342b2d83423a9790e84e82caacfd-d8001adad62e049f.gif</thumbnail_url><duration>324.766667</duration><title>Local RAG Wikipedia Assistant Setup and Tests</title><description>I built and tested a fully local RAG system with Olamas and a local Mistral model, ingesting a Wikipedia playlist of 40 famous people and places in about 2 to 3 minutes. I then ran an embedding store locally using a Mini LM model, saving a minicromovector based database. I demonstrated queries like what Mercury discovered, where Now is, and Albert Einstein questions, including chat memory and compression behavior. I also showed a no answer case with who is the president of Mars, returning I do not know. I showed the clear chat history button. No action was specifically requested from viewers.</description></oembed>