Calling a sub-workflow with a wait by webhook node, don't return the data of the last node, but the node before the wait.

[Arnaud-MARIE]

Describe the problem/error/question

When calling a sub-workflow containing a Wait (Webhook) node, the return data is incorrect:

• When called as Agent Tool: Returns data from node before Wait
• When called as Workflow: Returns initial input data
• Expected: Should return data from last node after webhook completion

Steps to Reproduce

1. Create sub-workflow with:

• Initial API call
• Wait (Webhook) node
• Final data processing

2. Call this sub-workflow from main workflow
3. Wait for webhook completion
4. Check returned data in main workflow

Current Behavior

• When called as Agent Tool: Returns data from the node before the Wait node
• When called as Workflow: Returns only the initial input data
• Webhook executes successfully but final data is not returned

Expected Behavior

Should return the processed data from the last node after webhook completion

Technical Details

• Using Wait node in 'Webhook' mode
• Webhook successfully triggers and completes
• Sub-workflow executes completely
• Issue occurs in data return to parent workflow

What is the error message (if any)?

Please share your workflow/screenshots/recording

Sub workflow : 
{
  "nodes": [
    {
      "parameters": {
        "resume": "webhook",
        "httpMethod": "POST",
        "responseMode": "lastNode",
        "limitWaitTime": true,
        "resumeAmount": 10,
        "resumeUnit": "minutes",
        "options": {}
      },
      "type": "n8n-nodes-base.wait",
      "typeVersion": 1.1,
      "position": [
        400,
        0
      ],
      "id": "ff89d4fa-5c3b-48ba-81cb-abb6554445ac",
      "name": "Wait response",
      "webhookId": "7e72cb0f-e29c-4e74-9509-95cc75e8e836"
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "e2347a4a-cc19-4049-8a19-e11a411b648d",
              "name": "test",
              "value": "={{ \n /(x|twitter)\\.com.*?status/.test($json.url) ? {\n   scraping_type: \"x post (ex-twitter)\",\n   dataset_id: \"gd_lwxkxvnf1cynvib9co\"\n } : \n /linkedin\\.com\\/posts/.test($json.url) ? {\n   scraping_type: \"linkedin post\", \n   dataset_id: \"gd_lyy3tktm25m4avu764\"\n } : null\n}}",
              "type": "object"
            }
          ]
        },
        "includeOtherFields": true,
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        -140,
        0
      ],
      "id": "bd71d82e-1b4e-4b52-b066-e3bf801331c4",
      "name": "Set variable based on social network"
    },
    {
      "parameters": {
        "method": "POST",
        "url": "https://api.brightdata.com/datasets/v3/trigger",
        "sendQuery": true,
        "queryParameters": {
          "parameters": [
            {
              "name": "dataset_id",
              "value": "={{ $json.test.dataset_id }}"
            },
            {
              "name": "include_errors",
              "value": "true"
            },
            {
              "name": "endpoint",
              "value": "={{  $execution.resumeUrl }}"
            },
            {
              "name": "format",
              "value": "json"
            },
            {
              "name": "uncompressed_webhook",
              "value": "true"
            },
            {
              "name": "force_deliver",
              "value": "true"
            }
          ]
        },
        "sendHeaders": true,
        "headerParameters": {
          "parameters": [
            {
              "name": "Authorization"
            }
          ]
        },
        "sendBody": true,
        "specifyBody": "json",
        "jsonBody": "=[\n  {\n    \"url\": \"{{ $json.url }}\"\n  }\n]",
        "options": {
          "response": {
            "response": {
              "fullResponse": true,
              "neverError": true
            }
          }
        }
      },
      "type": "n8n-nodes-base.httpRequest",
      "typeVersion": 4.2,
      "position": [
        180,
        0
      ],
      "id": "070df48f-adb7-4f8e-8c36-855a45f283c0",
      "name": "Scrap special dataset post1",
      "alwaysOutputData": true
    },
    {
      "parameters": {
        "assignments": {
          "assignments": [
            {
              "id": "2b1c7ff8-06a7-448b-99b7-5ede4b2e0bf0",
              "name": "response",
              "value": "={{ $json }}",
              "type": "object"
            }
          ]
        },
        "options": {}
      },
      "id": "ed06fc70-12fe-4d96-a135-8f20b3a6f566",
      "name": "Set response",
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.3,
      "position": [
        880,
        0
      ]
    },
    {
      "parameters": {
        "mode": "raw",
        "jsonOutput": "={\n  \"scraping_type\": \"{{ $('Set variable based on social network').first().json.test.scraping_type }}\",\n  \"username\": \"{{ $json.body[0].name || $json.body[0].headline }} \",\n  \"description\":{{ JSON.stringify($json.body[0].description?.replaceAll('\"',\"'\")) || JSON.stringify($json.body[0].post_text?.replaceAll('\"',\"'\"))}}\n}\n ",
        "options": {}
      },
      "type": "n8n-nodes-base.set",
      "typeVersion": 3.4,
      "position": [
        640,
        0
      ],
      "id": "21d7a234-7542-41be-b8cd-9d2f2dbee24a",
      "name": "Prepare Response"
    },
    {
      "parameters": {
        "inputSource": "passthrough"
      },
      "type": "n8n-nodes-base.executeWorkflowTrigger",
      "typeVersion": 1.1,
      "position": [
        -440,
        0
      ],
      "id": "63ec4196-1b1b-4dd0-af8f-4110abb6f833",
      "name": "When Executed by Another Workflow"
    }
  ],
  "connections": {
    "Wait response": {
      "main": [
        [
          {
            "node": "Prepare Response",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Set variable based on social network": {
      "main": [
        [
          {
            "node": "Scrap special dataset post1",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Scrap special dataset post1": {
      "main": [
        [
          {
            "node": "Wait response",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Set response": {
      "main": [
        []
      ]
    },
    "Prepare Response": {
      "main": [
        [
          {
            "node": "Set response",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "When Executed by Another Workflow": {
      "main": [
        [
          {
            "node": "Set variable based on social network",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  },
  "pinData": {
    "When Executed by Another Workflow": [
      {
        "url": "https://x.com/tom_doerr/status/1887628241587827097"
      }
    ]
  },
  "meta": {
    "templateCredsSetupCompleted": true,
    "instanceId": "8e95de061dd3893a50b8b4c150c8084a7848fb1df63f53533941b7c91a8ab996"
  }
}

Share the output returned by the last node

Parent Workflow :

[
  {
    "url": "https://x.com/tom_doerr/status/1887628241587827097"
  }
]

Sub-workflow :

[
  {
    "response": {
      "scraping_type": "x post (ex-twitter)",
      "username": "Tom Dörr ",
      "description": "Build AI agents with MCP"
    }
  }
]

Debug info

core

• n8nVersion: 1.77.3
• platform: docker (cloud)
• nodeJsVersion: 20.18.2
• database: sqlite
• executionMode: regular
• concurrency: 20
• license: community

storage

• success: all
• error: all
• progress: false
• manual: true
• binaryMode: filesystem

pruning

• enabled: true
• maxAge: 720 hours
• maxCount: 25000 executions

client

• userAgent: mozilla/5.0 (windows nt 10.0; win64; x64) applewebkit/537.36 (khtml, like gecko) chrome/132.0.0.0 safari/537.36
• isTouchDevice: false

Generated at: 2025-02-07T17:56:34.150Z}

[Joffcom]

Hey @Arnaud-MARIE,

We have created an internal ticket to look into this which we will be tracking as "GHC-768"

[Arnaud-MARIE]

Important Note

When replacing the Wait (Webhook) mode with a time-based Wait node (e.g., waiting for X seconds), the workflow returns the correct final data. This suggests the issue is specific to webhook-based waiting.

Here the modified workflow
{ "nodes": [ { "parameters": { "assignments": { "assignments": [ { "id": "e2347a4a-cc19-4049-8a19-e11a411b648d", "name": "test", "value": "={{ \n /(x|twitter)\\.com.*?status/.test($json.url) ? {\n scraping_type: \"x post (ex-twitter)\",\n dataset_id: \"gd_lwxkxvnf1cynvib9co\"\n } : \n /linkedin\\.com\\/posts/.test($json.url) ? {\n scraping_type: \"linkedin post\", \n dataset_id: \"gd_lyy3tktm25m4avu764\"\n } : null\n}}", "type": "object" } ] }, "includeOtherFields": true, "options": {} }, "type": "n8n-nodes-base.set", "typeVersion": 3.4, "position": [ -160, 240 ], "id": "bd71d82e-1b4e-4b52-b066-e3bf801331c4", "name": "Set variable based on social network" }, { "parameters": { "inputSource": "passthrough" }, "type": "n8n-nodes-base.executeWorkflowTrigger", "typeVersion": 1.1, "position": [ -460, 240 ], "id": "63ec4196-1b1b-4dd0-af8f-4110abb6f833", "name": "When Executed by Another Workflow" }, { "parameters": { "method": "POST", "url": "https://api.brightdata.com/datasets/v3/trigger", "sendQuery": true, "queryParameters": { "parameters": [ { "name": "dataset_id", "value": "={{ $json.test.dataset_id }}" }, { "name": "include_errors", "value": "true" }, { "name": "endpoint", "value": "={{ $execution.resumeUrl }}" }, { "name": "format", "value": "json" }, { "name": "uncompressed_webhook", "value": "true" }, { "name": "force_deliver", "value": "true" } ] }, "sendHeaders": true, "headerParameters": { "parameters": [ { "name": "Authorization" } ] }, "sendBody": true, "specifyBody": "json", "jsonBody": "=[\n {\n \"url\": \"{{ $json.url }}\"\n }\n]", "options": { "response": { "response": { "fullResponse": true, "neverError": true } } } }, "type": "n8n-nodes-base.httpRequest", "typeVersion": 4.2, "position": [ 200, 280 ], "id": "dde506e1-424b-43b9-9993-bd3f3d3763d8", "name": "Scrap special dataset post", "alwaysOutputData": true }, { "parameters": { "url": "=https://api.brightdata.com/datasets/v3/snapshot/{{ $json.body.snapshot_id }}", "sendQuery": true, "queryParameters": { "parameters": [ { "name": "format", "value": "json" } ] }, "sendHeaders": true, "headerParameters": { "parameters": [ { "name": "Authorization" } ] }, "options": {} }, "type": "n8n-nodes-base.httpRequest", "typeVersion": 4.2, "position": [ 760, 300 ], "id": "d55563c0-bb4c-4476-8efb-9e1756ed6225", "name": "Scrap special dataset post2", "alwaysOutputData": true }, { "parameters": { "assignments": { "assignments": [ { "id": "2b1c7ff8-06a7-448b-99b7-5ede4b2e0bf0", "name": "response", "value": "={{ $json }}", "type": "object" } ] }, "options": {} }, "id": "856ba58c-55be-4657-b39a-7b4c641b5c0d", "name": "Set response1", "type": "n8n-nodes-base.set", "typeVersion": 3.3, "position": [ 1860, 320 ] }, { "parameters": { "mode": "raw", "jsonOutput": "={\n \"scraping_type\": \"{{ $('Set variable based on social network').first().json.test.scraping_type }}\",\n \"username\": \"{{ $json.name || $json.headline }} \",\n \"description\":{{ JSON.stringify($json.description?.replaceAll('\"',\"'\")) || JSON.stringify($json.post_text?.replaceAll('\"',\"'\"))}}\n}\n", "options": {} }, "type": "n8n-nodes-base.set", "typeVersion": 3.4, "position": [ 1620, 320 ], "id": "dd819973-5019-49c7-8179-c266afcdbb01", "name": "Prepare Response1" }, { "parameters": { "amount": 10 }, "type": "n8n-nodes-base.wait", "typeVersion": 1.1, "position": [ 420, 280 ], "id": "8c9f6695-f171-4131-9e20-f6fa90dbbc49", "name": "Wait", "webhookId": "396472f7-b30b-44d3-a101-b3484c002ffa" }, { "parameters": { "rules": { "values": [ { "conditions": { "options": { "caseSensitive": true, "leftValue": "", "typeValidation": "strict", "version": 2 }, "conditions": [ { "leftValue": "={{ $json.status }}", "rightValue": "", "operator": { "type": "string", "operation": "exists", "singleValue": true } } ], "combinator": "and" }, "renameOutput": true, "outputKey": "Not ready" } ] }, "options": { "fallbackOutput": "extra" } }, "type": "n8n-nodes-base.switch", "typeVersion": 3.2, "position": [ 980, 300 ], "id": "ee0af8b7-04e1-4b33-b4ef-310a59b63292", "name": "Switch" } ], "connections": { "Set variable based on social network": { "main": [ [ { "node": "Scrap special dataset post", "type": "main", "index": 0 } ] ] }, "When Executed by Another Workflow": { "main": [ [ { "node": "Set variable based on social network", "type": "main", "index": 0 } ] ] }, "Scrap special dataset post": { "main": [ [ { "node": "Wait", "type": "main", "index": 0 } ] ] }, "Scrap special dataset post2": { "main": [ [ { "node": "Switch", "type": "main", "index": 0 } ] ] }, "Prepare Response1": { "main": [ [ { "node": "Set response1", "type": "main", "index": 0 } ] ] }, "Wait": { "main": [ [ { "node": "Scrap special dataset post2", "type": "main", "index": 0 } ] ] }, "Switch": { "main": [ [ { "node": "Wait", "type": "main", "index": 0 } ], [ { "node": "Prepare Response1", "type": "main", "index": 0 } ] ] } }, "pinData": { "When Executed by Another Workflow": [ { "url": "https://x.com/tom_doerr/status/1887628241587827097?t=O7GfOyVTAVIVgbT10dxeew&s=31" } ] }, "meta": { "templateCredsSetupCompleted": true, "instanceId": "8e95de061dd3893a50b8b4c150c8084a7848fb1df63f53533941b7c91a8ab996" } }

[astew]

I've observed the same as @Arnaud-MARIE

Additionally, I've found that if I simply hoist the contents of the sub-workflow out into the workflow that was calling it, then it all works as expected.

[brunoboto96]

I've noticed this also happens in this specific workflow:
https://n8n.io/workflows/2536-pattern-for-parallel-sub-workflow-execution-followed-by-wait-for-all-loop/

In my case, since I'm starting the workflow using a webhook and doing more stuff before and after this block, I can't send a response back.

I'm using this method to process multiple prompts at once.

A solution would be to have an identifier on the waiting node for which webhook response to wait for from a dropdown list. I believe then it wouldn't interfere with the main thread.

[TaipanRex]

This is an issue with all workflows that contain a asyncronous node, when the workflow is used as a sub-workflow. I'll give an example of the problem from one of my use cases: I have a workflow that uses Slack "send and await" operation to confirm with me some changes before executing them. If I run the workflow directly, this works fine. If I run it as a sub-workflow tool for a AI Agent, the Workflow Tool node output will always be the input to the Slack node in the sub-workflow. I think I know the reason why.

Look at the async execute code in packages/nodes-base/nodes/Slack/V2/SlackV2.node.ts:

if (resource === 'message' && operation === SEND_AND_WAIT_OPERATION) {
	await slackApiRequest.call(
		this,
		'POST',
		'/chat.postMessage',
		createSendAndWaitMessageBody(this),
	);

	const waitTill = configureWaitTillDate(this);

	await this.putExecutionToWait(waitTill);
	return [this.getInputData()];
}

The request is sent to Slack and the execution is put to wait (waitTill becomes whatever you had set on your Slack node as maximum time to wait). The Execution is successfully suspended, waiting either for the Slack callback or timing out. The problem is return [this.getInputData()];, the input data from the Slack node is returned from the Workflow tool to the AI Agent. The AI Agent will not get the final output after the execution is reactivated. This is basically an issue with any node that has this type of asyncronous behaviour.

The workaround I have been using is to instead use Webhook and Respond to Webhook nodes, using a HTTP call tool instead of Workflow Tool for my agent. For some reason this removes the async issue and the agent will wait until the Respond to Webhook node completes. Its not ideal because the output of that node is always a {"response": "your output stringified"} object instead of your actual output object.

[cyranius]

+1

[geraldo-augusto-jr]

+1

[Cadiac]

Thanks for the report, this reproducible and the issue is being investigated. This also affects Wait nodes in After Time Interval mode if the wait timer is longer than 65 seconds.

[DimonHo]

+1
Additionally, may I ask, in this process, it seems that the two subprocesses are not executed in parallel?