AI / LangGraph LangChain Interview questions
How do you implement human-in-the-loop with LangGraph?
Human-in-the-loop (HITL) in LangGraph means pausing graph execution at a specified point so a human can inspect the current state, approve an action, or modify a value before the graph continues. This is a first-class LangGraph feature built on top of checkpointing.
Step 1: Compile the graph with interrupt_before or interrupt_after
from langgraph.checkpoint.memory import MemorySaver graph = graph_builder.compile( checkpointer=MemorySaver(), interrupt_before=["tool_executor"], # pause before this node runs )
Step 2: Run until the interrupt
config = {"configurable": {"thread_id": "session-1"}} for event in graph.stream({"messages": [HumanMessage("Search for X")]}, config): print(event) # stops before tool_executor
Step 3: Inspect and optionally update state
current_state = graph.get_state(config) print(current_state.values) # see what the agent is about to do # Optionally modify the state before continuing: graph.update_state(config, {"messages": [HumanMessage("Actually search for Y")]}, as_node="agent")
Step 4: Resume execution
# Pass None as input to resume from the checkpoint for event in graph.stream(None, config): print(event)
What argument to graph.compile() causes execution to pause before a specific node?
How do you resume a LangGraph run after a human-in-the-loop pause?
More Related questions...
