will probably need to custom code it. I'm sure you could use the server url updates to gather context and send it to an LLM every x seconds though. Any thoughts? @Chowderr