N8n-nodes-query-retriever-rerank stopped working

Describe the problem/error/question

At some point after the update of n8n docker to latest the workflow stopped working. n8n-nodes-query-retriever-rerank produces an error. @blah any ideas as an author of a node?

What is the error message (if any)?

{
“errorMessage”: “Invalid URL”,
“errorDetails”: {
“rawErrorMessage”: [
“Invalid URL”
],
“httpCode”: “none”
},
“n8nDetails”: {
“nodeName”: “KB”,
“nodeType”: “n8n-nodes-query-retriever-rerank.queryRetrieverRerank”,
“nodeVersion”: 1,
“itemIndex”: 0,
“runIndex”: 0,
“time”: “1/23/2026, 3:51:37 PM”,
“n8nVersion”: “2.4.5 (Self Hosted)”,
“binaryDataMode”: “filesystem”,
“stackTrace”: [
“NodeApiError: Invalid URL”,
" at RoutingNode.runNode (/usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_ec37920eb95917b28efaa783206b20f3/node_modules/n8n-core/src/execution-engine/routing-node.ts:258:11)“,
" at processTicksAndRejections (node:internal/process/task_queues:105:5)”,
" at ExecuteContext.versionedNodeType.execute (/usr/local/lib/node_modules/n8n/src/node-types.ts:60:18)“,
" at WorkflowExecute.executeNode (/usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_ec37920eb95917b28efaa783206b20f3/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:1045:8)”,
" at WorkflowExecute.runNode (/usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_ec37920eb95917b28efaa783206b20f3/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:1226:11)“,
" at /usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_ec37920eb95917b28efaa783206b20f3/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:1662:27”,
" at /usr/local/lib/node_modules/n8n/node_modules/.pnpm/n8n-core@file+packages+core_@[email protected]_@[email protected]_ec37920eb95917b28efaa783206b20f3/node_modules/n8n-core/src/execution-engine/workflow-execute.ts:2297:11"
]
}
}

Please share your workflow

{“nodes”: [{“parameters”: {“tableName”: “n8n_ai_document”,“options”: {}},“type”: “/n8n-nodes-langchain.vectorStorePGVector”,“typeVersion”: 1.1,“position”: [352,528],“id”: “d3cae887-5440-4d99-b832-52fbd6137a82”,“name”: “Vector store”,“retryOnFail”: true,“credentials”: {“postgres”: {“id”: “VaPYxbBQRMKprMA8”,“name”: “Postgres account”}}},{“parameters”: {“promptType”: “define”,“text”: “=Overview\nYou are a personal assistant who has access to documents on different topics.\n\n## Tools\n"Think" - Use this tool to find solution when you run into an issue before asking the user.\n"knowledge_base" - Use this tool to search in the "n8n_ai_document" main documents table.\n"Calculator" - Use it to calculate results for mathematical problems.\n"Fallback query documents" - Use this tool as fallback to get documents from "knowledge_base" with sql query filtered by subjects or keywords.\n\n## Instructions\n1) The question from the user is: {{ $json.enhanced_message }}\n2) Start by performing search in "knowledge_base" by using the part of the question within <kb_query></kb_query> tags without these tags.\n3) Use the "Think" tool to verify you took the right steps before answering.\n4) Use "Fallback query documents" if until here the answer is negative.\n\n## Important\nUse "Think" tool and then "Fallback query documents" before giving final answer.\n\n/think”,“options”: {“enableStreaming”: false}},“type”: “/n8n-nodes-langchain.agent”,“typeVersion”: 3,“position”: [464,96],“id”: “c00174df-f36d-40af-a63a-fb3a0b212738”,“name”: “AI Agent1”,“retryOnFail”: true},{“parameters”: {“model”: “Qwen3-Embedding-8B-f16.gguf”,“options”: {}},“type”: “/n8n-nodes-langchain.embeddingsOpenAi”,“typeVersion”: 1.2,“position”: [704,640],“id”: “26833bf8-b7ab-4454-a208-70e6c977bb31”,“name”: “Llama_Qwen_Embedding”,“credentials”: {“openAiApi”: {“id”: “wFta8QMBmblNsoED”,“name”: “Llama_Qwen_Embedding”}}},{“parameters”: {“model”: {“__rl”: true,“value”: “Qwen3-8B-UD-Q4_K_XL.gguf”,“mode”: “list”,“cachedResultName”: “Qwen3-8B-UD-Q4_K_XL.gguf”},“builtInTools”: {},“options”: {“timeout”: 600000}},“type”: “/n8n-nodes-langchain.lmChatOpenAi”,“typeVersion”: 1.3,“position”: [576,656],“id”: “57ea1dcd-891e-47c9-8503-8b223333babf”,“name”: “Llama_Qwen3_reasoning1”,“credentials”: {“openAiApi”: {“id”: “6NxSjYgOGYSzzxlD”,“name”: “LlamaCpp_Qwen3”}}},{“parameters”: {“toolOptions”: {“toolName”: “knowledge_base”},“retrievalOptions”: {},“queryStrategy”: {“strategyType”: “simple_query”},“options”: {}},“type”: “n8n-nodes-query-retriever-rerank.queryRetrieverRerank”,“typeVersion”: 1,“position”: [512,352],“id”: “f41983d8-8b39-4317-b61a-64f24e92136a”,“name”: “KB”}],“connections”: {“Vector store”: {“ai_vectorStore”: [[{“node”: “KB”,“type”: “ai_vectorStore”,“index”: 0}]]},“AI Agent1”: {“main”: []},“Llama_Qwen_Embedding”: {“ai_embedding”: [[{“node”: “Vector store”,“type”: “ai_embedding”,“index”: 0},{“node”: “KB”,“type”: “ai_embedding”,“index”: 0}]]},“Llama_Qwen3_reasoning1”: {“ai_languageModel”: [[{“node”: “KB”,“type”: “ai_languageModel”,“index”: 0}]]},“KB”: {“ai_tool”: [[{“node”: “AI Agent1”,“type”: “ai_tool”,“index”: 0}]]}},“pinData”: {},“meta”: {“templateCredsSetupCompleted”: true,“instanceId”: “69591f51bb014cc8635030b0a7318ef5d586c6bd856016428c998902fe6afe12”}}

Share the output returned by the last node

Information on your n8n setup

  • n8n version: 2.4.5
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker):
  • Operating system: Windows 11

I checked my workflow with older version of n8n docker and it worked there. I found out, that it had older version of the AI_Agent node, which I replaced with the new version of the node in newest n8n and forgot about it, so it looks like with the older version of the AI_Agent node in newest n8n the workflow still works fine. It’s newest AI_Agent node version, that has an issue with communicating with retriever rerank node. Any chance to fix the bug?

Version of the AI_Agent node, that works is 1.9 and the one that does not work is 3.1