Skip to content

How to wait for user input using interrupt

Prerequisites

This guide assumes familiarity with the following concepts:

Human-in-the-loop (HIL) interactions are crucial for agentic systems. Waiting for human input is a common HIL interaction pattern, allowing the agent to ask the user clarifying questions and await input before proceeding.

We can implement this in LangGraph using the function. interrupt allows us to stop graph execution to collect input from a user and continue execution with collected input.

Setup

We are not going to show the full code for the graph we are hosting, but you can see it here if you want to. Once this graph is hosted, we are ready to invoke it and wait for user input.

SDK initialization

First, we need to setup our client so that we can communicate with our hosted graph:

from langgraph_sdk import get_client
client = get_client(url=<DEPLOYMENT_URL>)
# Using the graph deployed with the name "agent"
assistant_id = "agent"
thread = await client.threads.create()

import { Client } from "@langchain/langgraph-sdk";

const client = new Client({ apiUrl: <DEPLOYMENT_URL> });
// Using the graph deployed with the name "agent"
const assistantId = "agent";
const thread = await client.threads.create();

curl --request POST \
  --url <DEPLOYMENT_URL>/threads \
  --header 'Content-Type: application/json' \
  --data '{}'

Waiting for user input

Initial invocation

Now, let's invoke our graph.

input = {
    "messages": [
        {
            "role": "user",
            "content": "Ask the user where they are, then look up the weather there",
        }
    ]
}

async for chunk in client.runs.stream(
    thread["thread_id"],
    assistant_id,
    input=input,
    stream_mode="updates",
):
    if chunk.data and chunk.event != "metadata": 
        print(chunk.data)

const input = {
  messages: [
    {
      role: "human",
      content: "Ask the user where they are, then look up the weather there" }
  ]
};

const streamResponse = client.runs.stream(
  thread["thread_id"],
  assistantId,
  {
    input: input,
    streamMode: "updates",
  }
);

for await (const chunk of streamResponse) {
  if (chunk.data && chunk.event !== "metadata") {
    console.log(chunk.data);
  }
}

curl --request POST \
 --url <DEPLOYMENT_URL>/threads/<THREAD_ID>/runs/stream \
 --header 'Content-Type: application/json' \
 --data "{
   \"assistant_id\": \"agent\",
   \"input\": {\"messages\": [{\"role\": \"human\", \"content\": \"Ask the user where they are, then look up the weather there\"}]},
   \"stream_mode\": [
     \"updates\"
   ]
 }"

Output:

{'agent': {'messages': [{'content': [{'text': "I'll help you ask the user about their location and then search for weather information.", 'type': 'text'}, {'id': 'toolu_012JeNEvyePZFWK39d52Wdwi', 'input': {'question': 'Where are you located?'}, 'name': 'AskHuman', 'type': 'tool_use'}], 'additional_kwargs': {}, 'response_metadata': {'id': 'msg_01UBEdS6UvuFMetdokNsykVG', 'model': 'claude-3-5-sonnet-20241022', 'stop_reason': 'tool_use', 'stop_sequence': None, 'usage': {'cache_creation_input_tokens': 0, 'cache_read_input_tokens': 0, 'input_tokens': 438, 'output_tokens': 76}, 'model_name': 'claude-3-5-sonnet-20241022'}, 'type': 'ai', 'name': None, 'id': 'run-1b1210d8-39e0-4607-9f0e-0ea932d28d5c-0', 'example': False, 'tool_calls': [{'name': 'AskHuman', 'args': {'question': 'Where are you located?'}, 'id': 'toolu_012JeNEvyePZFWK39d52Wdwi', 'type': 'tool_call'}], 'invalid_tool_calls': [], 'usage_metadata': {'input_tokens': 438, 'output_tokens': 76, 'total_tokens': 514, 'input_token_details': {'cache_read': 0, 'cache_creation': 0}}}]}}
{'__interrupt__': [{'value': 'Where are you located?', 'resumable': True, 'ns': ['ask_human:2d41f894-f297-211e-9bfe-1d162ecba54a'], 'when': 'during'}]}

You can see that our graph got interrupted inside the ask_human node, which is now waiting for a location to be provided.

Providing human input

We can provide human input (location) by invoking the graph with a Command(resume="<location>"):

from langgraph_sdk.schema import Command

async for chunk in client.runs.stream(
    thread["thread_id"],
    assistant_id,
    command=Command(resume="san francisco"),
    stream_mode="updates",
):
    if chunk.data and chunk.event != "metadata": 
        print(chunk.data)

const streamResponse = client.runs.stream(
  thread["thread_id"],
  assistantId,
  {
    command: { resume: "san francisco" },
    streamMode: "updates"
  }
);

for await (const chunk of streamResponse) {
  if (chunk.data && chunk.event !== "metadata") {
    console.log(chunk.data);
  }
}

curl --request POST \
 --url <DEPLOYMENT_URL>/threads/<THREAD_ID>/runs/stream \
 --header 'Content-Type: application/json' \
 --data "{
   \"assistant_id\": \"agent\",
   \"command\": {
     \"resume\": \"san francisco\"
   },
   \"stream_mode\": [
     \"updates\"
   ]
 }"

Output:

{'ask_human': {'messages': [{'tool_call_id': 'toolu_012JeNEvyePZFWK39d52Wdwi', 'type': 'tool', 'content': 'san francisco'}]}}
{'agent': {'messages': [{'content': [{'text': 'Let me search for the weather in San Francisco.', 'type': 'text'}, {'id': 'toolu_019f9Y7ST6rNeDQkDjFCHk6C', 'input': {'query': 'current weather in san francisco'}, 'name': 'search', 'type': 'tool_use'}], 'additional_kwargs': {}, 'response_metadata': {'id': 'msg_0152YFm7DtnzfZQuiMUzaSsw', 'model': 'claude-3-5-sonnet-20241022', 'stop_reason': 'tool_use', 'stop_sequence': None, 'usage': {'cache_creation_input_tokens': 0, 'cache_read_input_tokens': 0, 'input_tokens': 527, 'output_tokens': 67}, 'model_name': 'claude-3-5-sonnet-20241022'}, 'type': 'ai', 'name': None, 'id': 'run-f509b5b2-eb30-4200-a8da-fa79ed68812a-0', 'example': False, 'tool_calls': [{'name': 'search', 'args': {'query': 'current weather in san francisco'}, 'id': 'toolu_019f9Y7ST6rNeDQkDjFCHk6C', 'type': 'tool_call'}], 'invalid_tool_calls': [], 'usage_metadata': {'input_tokens': 527, 'output_tokens': 67, 'total_tokens': 594, 'input_token_details': {'cache_read': 0, 'cache_creation': 0}}}]}}
{'action': {'messages': [{'content': "I looked up: current weather in san francisco. Result: It's sunny in San Francisco, but you better look out if you're a Gemini 😈.", 'additional_kwargs': {}, 'response_metadata': {}, 'type': 'tool', 'name': 'search', 'id': 'cbd0f623-cc12-48a2-8c18-3cbb943e46e0', 'tool_call_id': 'toolu_019f9Y7ST6rNeDQkDjFCHk6C', 'artifact': None, 'status': 'success'}]}}
{'agent': {'messages': [{'content': "Based on the search results, it's currently sunny in San Francisco. Would you like any specific details about the weather forecast?", 'additional_kwargs': {}, 'response_metadata': {'id': 'msg_01FhzXj72CehBYkJGX69vsBc', 'model': 'claude-3-5-sonnet-20241022', 'stop_reason': 'end_turn', 'stop_sequence': None, 'usage': {'cache_creation_input_tokens': 0, 'cache_read_input_tokens': 0, 'input_tokens': 639, 'output_tokens': 29}, 'model_name': 'claude-3-5-sonnet-20241022'}, 'type': 'ai', 'name': None, 'id': 'run-f48e818e-dd88-415e-9a0b-4a958498b553-0', 'example': False, 'tool_calls': [], 'invalid_tool_calls': [], 'usage_metadata': {'input_tokens': 639, 'output_tokens': 29, 'total_tokens': 668, 'input_token_details': {'cache_read': 0, 'cache_creation': 0}}}]}}