<?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/f82cc412ed86454c88e10b8ee1910752&quot; frameborder=&quot;0&quot; width=&quot;1366&quot; height=&quot;1024&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1024</height><width>1366</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1024</thumbnail_height><thumbnail_width>1366</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/f82cc412ed86454c88e10b8ee1910752-d798901e0668afb3.gif</thumbnail_url><duration>297.862</duration><title>Automated Amazon Product Roast Generator Demo 🎬</title><description>Hi everyone, I am Sai Sathvik. In this Loom, I show my automated AI influencer list that generates a roast video from Amazon product data. I explain how I handle messy Amazon product layouts with a Melty APA fallback architecture so pipelines do not crash. The backend uses SERP APIK for deep Amazon scanning, a Grok LLM with Llama 3 style generation for the roast scripts, 11labs for high fidelity voice, and a movie pipeline for vehicle images. I built a Streamlit front end with a Generate video button, so you can try searching for products like watches and iPhones.</description></oembed>