<?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/dc330a59c16b4e668422a9a193f9a11a&quot; frameborder=&quot;0&quot; width=&quot;1320&quot; height=&quot;990&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>990</height><width>1320</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>990</thumbnail_height><thumbnail_width>1320</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/dc330a59c16b4e668422a9a193f9a11a-d9a58cc816631ffb.gif</thumbnail_url><duration>327.531</duration><title>SoundMatch Agentic AI Music Recommender Demo</title><description>Hi, I am Sean Clemens. I built SoundMatch, an agentic music recommender that includes retrieval, self critique, confidence scoring, and automated testing, plus a UI that shows all five steps in real time. With default settings, Sunrise City ranked highest at 1.0 confidence, high, and I show the exact genre and mood reasons for each pick. When I changed mode to mood first, confidence rose to 94 percent. I then tested guardrails where an out of range energy of 1.5 was blocked, and my test harness ran 8 out of 8 tests passed. No action was requested from viewers.</description></oembed>