LangGraph adds built-in error handling for long-running AI agents

Original: lots of things can go wrong when you have long running agents, including model/tool call failures and rate limits

Source: x.com ↗