AI Agents & LangChain in n8n · 40 min

Guardrails, evals & testing agents

Regex filters, LLM-as-judge scoring, and repeatable eval workflows before shipping an agent.

Key concepts

  • Input/output guardrails
  • LLM-as-judge pattern
  • Golden-set regression testing

Exercise

Build a 10-question golden set and an eval workflow that scores each agent response 1-5.

Official docs