<?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/01fb7c0289d8420f8ae724aad0a6d165&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/01fb7c0289d8420f8ae724aad0a6d165-f598355c8087952f.gif</thumbnail_url><duration>1405.408</duration><title>GUI Agent_1</title><description>This Loom demos a GUI automation agent architecture that logs into the Higgs field platform using a React loop and multi-modal perception. It uses Playwright as the browser and input tool, with OmniParser performing screenshot annotation and OCR, then a Qwen model deciding the next bounding box and whether keyboard input is needed across human-in-the-loop levels. The agent navigates sign-in options, uses a fast-path for safe back or login flows, and requests password and OTP when required for security in a Chromium session. It saves execution logs and ends with a success flag after successful login, with screenshots and reasoning shown during the run.</description></oembed>