Basic LLM Chain fails with "asyncIterator" error on Cloud v1.118.1 despite perfect input

Hi n8n community,

I’m running into a persistent error with the Basic LLM Chain node and am hoping someone can help. The node consistently fails with an asyncIterator error, even when fed a perfectly formatted, static input.

The Problem

My workflow is designed to enrich Reddit posts with Google News and then use the Basic LLM Chain node to generate a video plan. However, the Basic LLM Chain node always fails with the following error message:

NodeOperationError: Cannot read properties of undefined (reading ‘Symbol(Symbol.asyncIterator)’)

Key Details:

  • n8n Version: 1.118.1 (Cloud)

  • Affected Nodes:

    • Basic LLM Chain (@n8n/n8n-nodes-langchain.chainLlm)

    • Google Vertex Chat Model (@n8n/n8n-nodes-langchain.lmChatGoogleVertex)

    • Structured Output Parser (@n8n/n8n-nodes-langchain.outputParserStructured)

What we’ve already tried

We have taken a full range of troubleshooting steps to ensure the input is correct:

  1. Corrected Data Flow: The workflow has been rebuilt using the standard “enrichment” pattern (Set → Branch → Google News → Merge) to prevent data loss.

  2. Filtering: The flow correctly filters duplicates and uses an IF node to stop cleanly when there are no new items to process.

  3. Robust Prompt: The prompt inside the Basic LLM Chain uses fallback values (e.g., {{ $json.prompt || “Default” }}) to handle potentially missing data.

The Decisive Test:
We isolated the problem by disconnecting the entire upstream workflow and using a single Code node to send a static, 100% correctly formatted item to the Basic LLM Chain. The error still occurs.

This proves that the issue is not with the incoming data, but likely within the node itself or in the interaction between the LangChain nodes.

Reproducible Example

Here is the exact data that reproduces the error:

Input to the Basic LLM Chain node (from the test Code node):

codeJSON

[
  {
    "json": {
      "prompt": "Test-Nachricht über den Tod von James Watson",
      "reddit_title": "Das ist ein Test-Titel von Reddit",
      "news_title": "- News-Titel 1\n- News-Titel 2",
      "news_url": "http://example.com"
    }
  }
]

Error Output from the Basic LLM Chain node:

codeJSON

{
  "errorMessage": "Cannot read properties of undefined (reading 'Symbol(Symbol.asyncIterator)')",
  "errorDetails": {
    "rawErrorMessage": [
      "Cannot read properties of undefined (reading 'Symbol(Symbol.asyncIterator)')"
    ]
  },
  "n8nDetails": {
    "nodeName": "Basic LLM Chain",
    "nodeType": "@n8n/n8n-nodes-langchain.chainLlm",
    "nodeVersion": 1.7,
    "time": "8.11.2025, 20:10:25",
    "n8nVersion": "1.118.1 (Cloud)",
    "binaryDataMode": "filesystem",
    "stackTrace": [
      "NodeOperationError: Cannot read properties of undefined (reading 'Symbol(Symbol.asyncIterator)')",
      "    at /usr/local/lib/node_modules/n8n/node_modules/.pnpm/@n8n+n8n-nodes-langchain@file+packages+@n8n+nodes-langchain_864cf0c09dffe523d02bd5c66425efbf/node_modules/@n8n/n8n-nodes-langchain/nodes/chains/ChainLLM/ChainLlm.node.ts:107:13",
      "    at Array.forEach (<anonymous>)",
      "    at ExecuteContext.execute (/usr/local/lib/node_modules/n8n/node_modules/.pnpm/@n8n+n8n-nodes-langchain@file+packages+@n8n+nodes-langchain_864cf0c09dffe523d02bd5c66425efbf/node_modules/@n8n/n8n-nodes-langchain/nodes/chains/ChainLLM/ChainLlm.node.ts:85:18)",
      "    at processTicksAndRejections (node:internal/process/task_queues:105:5)",
      "    at WorkflowExecute.executeNode (/usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_08b575bec2313d5d8a4cc75358971443/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:1093:8)",
      "    at WorkflowExecute.runNode (/usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_08b575bec2313d5d8a4cc75358971443/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:1274:11)",
      "    at /usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_08b575bec2313d5d8a4cc75358971443/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:1708:27",
      "    at /usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_08b575bec2313d5d8a4cc75358971443/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:2324:11"
    ]
  }
}

My Question

Since the error occurs even with a perfect, static input, I suspect this might be an internal bug in the LangChain nodes for this version of n8n.

Has anyone else encountered this problem? Is there a known workaround, or is this a case for the engineering team?

Thanks for any help or insight you can provide

See my reply in this topic (it won’t let me repeat myself in multiple topics): Vertex AI LangChain Agent fails ('asyncIterator' error) when Postgres Chat Memory is connected

I have also presented the same problem but I have the n8n self hosted version.