{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/d4ee837fba8e4b809cb8760c00c125fd\" frameborder=\"0\" width=\"1108\" height=\"831\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":831,"width":1108,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":831,"thumbnail_width":1108,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/d4ee837fba8e4b809cb8760c00c125fd-5f5fb721bc64a1f0.gif","duration":885.971,"title":"DreamUp QA - Demo","description":"In this video, I provide a status update on the DreamUp QA project, which is an automated browser testing agent for HTML5 games. I walk through the architecture of the agent, explaining how it interacts with a cloud-based Chrome browser and utilizes a large language model (LLM) to analyze gameplay. We have three testing modes: a normal LLM mode, a pause mode for games with accessible source code, and a quick test mode for basic functionality checks. I also discuss the evaluation phase, where we generate reports based on gameplay metrics and LLM reasoning. Please review the sample commands and configuration options provided, as your feedback will be valuable for refining our approach."}