Introducing LangSmith Engine

Most teams building agents today have traces. What they don't have is a clear answer to the question that actually matters: how should I improve my agent? Today, the process is by hand. You comb through traces to find errors, dig in to figure out what went wrong, ship a fix, and maybe write a test case to ensure it doesn’t regress something else. LangSmith Engine investigates your traces, paying close attention to any explicit errors, online eval failures, negative user feedback, or new behaviors the agent doesn't handle well yet. When it spots a problem, it looks for the same pattern across your project and clusters everything into one issue. Here's a quick explanation from LangChain CEO Harrison Chase and Product Manager Ben Tannyhill. What is LangSmith Engine? 00:00 Introduction to LangSmith Engine 00:15 The manual process of agent improvement 00:35 How LangSmith Engine works 00:56 Issue triage and PR generation 01:19 Custom evaluators and regression prevention 01:43 Benefits and real-world results Resources: Blog: https://www.langchain.com/blog/introducing-engine LangSmith: https://www.langchain.com/langsmith-platform LangChain: https://www.langchain.com/ https://youtu.be/L4pDS1uHl9o