{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/be0534bc818d408bab937033c6457ec9\" frameborder=\"0\" width=\"1872\" height=\"1404\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1404,"width":1872,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1404,"thumbnail_width":1872,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/be0534bc818d408bab937033c6457ec9-aa62dc3b25ed35b4.gif","duration":149.589,"title":"SIA setup and running task","description":"This Loom explains how to set up and run CR, including installing dependencies and configuring environment variables. The speaker creates a Python virtual environment supporting Python 3.11 to 3.14, then installs GID, which includes everything unless using an open source model that also requires installing OpenHands. To run a task, they note that CR includes baked-in tasks or you can pass your own task, and they mention using action as a chain code with steps to improve score and then run ID. They also stress setting the required environment variable for the base metamodel, where an Anthropy key is needed, and they preview how to bring your own agent profile via JSON by specifying the agent scaffold, provider details like endpoint and API key environment variable, and model information."}