Something we've been working on for the past few weeks - an easy way to visualize, debug, and explore the execution trace of LangChain agents and chains