Skip to content
masterclass Buddy
Start learning

AI Agents · Concept

Human-in-the-Loop

LangChain pauses selected tool calls for human review before execution. A reviewer can approve, edit, reject, or answer a call, and checkpointing lets the agent resume later.

Last updated

Syntax

HumanInTheLoopMiddleware(interrupt_on, description_prefix="Tool execution requires approval", interrupt_mode="batched", edit_notice=None)

In-depth API docs are only for subscribers: parameters, examples, common mistakes and interview practice.

Subscribe
  • HumanInTheLoopMiddleware
  • interrupt_on
  • interrupt_mode
  • Command
  • LangGraph persistence
  • InMemorySaver
  • action_requests
  • review_configs

Source: Human-in-the-loop. Checked 2026-11-03.