My property somehow is stated as “undefined”. But you can see on the left corner it does exist.
The whole flow is a single line:
What makes it impossible to be retrieved? How do I work around it?
Here is my workflow JSON:
{
“nodes”: [
{
“parameters”: {
“options”: {}
},
“type”: “@n8n/n8n-nodes-langchain.embeddingsOpenAi”,
“typeVersion”: 1.2,
“position”: [
-2380,
880
],
“id”: “42364de5-126c-4415-b130-cf829cbffa9e”,
“name”: “Embeddings OpenAI2”,
“credentials”: {
“openAiApi”: {
“id”: “he7mc0Z8YQRxDH1e”,
“name”: “OpenAi account”
}
}
},
{
“parameters”: {
“model”: {
“__rl”: true,
“mode”: “list”,
“value”: “gpt-4o-mini”
},
“options”: {}
},
“type”: “@n8n/n8n-nodes-langchain.lmChatOpenAi”,
“typeVersion”: 1.2,
“position”: [
-2340,
580
],
“id”: “4a7e0752-fde0-408b-a315-ed67ca0880b4”,
“name”: “OpenAI Chat Model1”,
“credentials”: {
“openAiApi”: {
“id”: “he7mc0Z8YQRxDH1e”,
“name”: “OpenAi account”
}
}
},
{
“parameters”: {
“modelId”: {
“__rl”: true,
“value”: “gpt-4o-mini”,
“mode”: “list”,
“cachedResultName”: “GPT-4O-MINI”
},
“messages”: {
“values”: [
{
“content”: “=You are Javan Zhang’s AI clone - 19yo entrepreneur specializing in systemized scaling. Use n8n outputs for all responses.\n\nCore Protocol:\n1. Response Style: \n - Lead with signature phrase (Yessir!/Scaling with systems baby)\n - 1 personal anecdote per interaction \n - Respond Conversationally and ensure most relevant response\n\n2. Value Stack:\n • Automation-first solutions\n • Bottleneck identification \n • Growth mindset triggers\n\n3. Phrase Matrix: \n High Frequency: "Let me break it down", "This changes everything…" \n Moderate: "Data shows…", "Prototype this tonight"\n\nEfficiency Rules:\n→ If query > 15 words: "Before we dive deep - crystal clear on your goal?"\n→ Code responses as [Action]: [Reason] format\n→ Never exceed 3 consecutive sentences\n”,
“role”: “system”
},
{
“content”: “=Your goal is to get all the data that is being provided and tailor/customize it specifically around the user query so it’s 100% relevant. Respond as Javan Zhang would\n\nThe response should be one single blend of information as a coach would put it”
},
{
“content”: “=User Query: {{ $(‘Body Text’).item.json.text }}\n\nData:\n\n{{ $json.takeaways }}”
}
]
},
“options”: {}
},
“type”: “@n8n/n8n-nodes-langchain.openAi”,
“typeVersion”: 1.8,
“position”: [
-1520,
340
],
“id”: “90752b42-024b-4d0d-8832-290433f6ba23”,
“name”: “OpenAI1”,
“credentials”: {
“openAiApi”: {
“id”: “he7mc0Z8YQRxDH1e”,
“name”: “OpenAi account”
}
}
},
{
“parameters”: {
“jsonSchemaExample”: “[{\n "videoTitle": "Building Coaching Programs That Scale",\n "link": "https://youtu.be/abc123\”,\n "timeStamp": "[03:28:50 - 03:35:53]",\n "takeaways": "Structure programs around client pain points, price based on outcomes, and validate demand through pre-launch surveys."\n}]"
},
“type”: “@n8n/n8n-nodes-langchain.outputParserStructured”,
“typeVersion”: 1.2,
“position”: [
-1960,
560
],
“id”: “216fb4e8-3b43-42e0-b5f8-4432b5f3b2c7”,
“name”: “Structured Output Parser”
},
{
“parameters”: {
“promptType”: “define”,
“text”: “=Query: {{ $json.text }}”,
“hasOutputParser”: true,
“options”: {
“systemMessage”: "=AI assistant retrieving and structuring YouTube transcript snippets with timestamped links",\n Find relevant transcript snippets matching query parameters Combine segments within 40s overlap/adjacency",\n Merge semantically related content\n Extend timestamp start -5-15s/end +5-15s (within video bounds)\nsummarize\ntakeaway_requirements\n Specific Actionable steps (35-60 words)",\n Clear value proposition\n Omit descriptive fluff\n Concise actionable summary\nMaximum token efficiency response\n timeStamp formatting: "[03:28:50 - 03:35:53]",\n\n "
}
},
“type”: “@n8n/n8n-nodes-langchain.agent”,
“typeVersion”: 1.7,
“position”: [
-2280,
340
],
“id”: “01f405ff-47f4-49a1-8c44-9643f5f59262”,
“name”: “Deepseeked De-Jsoned”
},
{
“parameters”: {
“jsCode”: “// Extract and process input items\nconst items = $input.all();\n\nconst groupedVideos = {};\n\n// Process each video entry\nitems.forEach(item => {\n item.json.output.forEach(video => {\n // Extract the first timestamp (HH:MM:SS)\n const match = video.timeStamp.match(/\[(\d{2}):(\d{2}):(\d{2})/);\n if (!match) return; // Skip if no valid timestamp\n\n // Convert timestamp to total seconds\n const hours = parseInt(match[1], 10);\n const minutes = parseInt(match[2], 10);\n const seconds = parseInt(match[3], 10);\n const startSeconds = hours * 3600 + minutes * 60 + seconds;\n\n // Identify video group by unique key (title + base link)\n const videoKey = ${video.videoTitle}_${video.link.split('&t=')[0]};\n\n if (!groupedVideos[videoKey]) {\n groupedVideos[videoKey] = ;\n }\n\n // Store video data with computed timestamp\n groupedVideos[videoKey].push({\n …video,\n startSeconds\n });\n });\n});\n\n// Function to merge timestamps and keep all takeaways\nconst mergeTimeRanges = (videos) => {\n videos.sort((a, b) => a.startSeconds - b.startSeconds);\n\n const mergedVideos = ;\n let currentMerge = { …videos[0], takeaways: [videos[0].takeaways] };\n\n for (let i = 1; i < videos.length; i++) {\n const currentVideo = videos[i];\n\n // Extract end time of the current merged segment\n const endTimeMatch = currentMerge.timeStamp.match(/\[(\d{2}):(\d{2}):(\d{2}) - (\d{2}):(\d{2}):(\d{2})]/);\n let endSeconds = currentMerge.startSeconds;\n \n if (endTimeMatch) {\n const endHours = parseInt(endTimeMatch[4], 10);\n const endMinutes = parseInt(endTimeMatch[5], 10);\n const endSecs = parseInt(endTimeMatch[6], 10);\n endSeconds = endHours * 3600 + endMinutes * 60 + endSecs;\n }\n\n // If the next segment is within 6 minutes, merge it\n if (currentVideo.startSeconds - endSeconds <= 360) {\n const newEndTime = currentVideo.timeStamp.match(/ - (\d{2}):(\d{2}):(\d{2})]/);\n if (newEndTime) {\n currentMerge.timeStamp = currentMerge.timeStamp.replace(/ - \d{2}:\d{2}:\d{2}/, - ${newEndTime[1]}:${newEndTime[2]}:${newEndTime[3]});\n }\n // Append takeaways instead of replacing\n currentMerge.takeaways.push(currentVideo.takeaways);\n } else {\n // Push the completed merge and start a new one\n currentMerge.takeaways = currentMerge.takeaways.join(" "); // Convert array to string\n mergedVideos.push(currentMerge);\n currentMerge = { …currentVideo, takeaways: [currentVideo.takeaways] };\n }\n }\n\n currentMerge.takeaways = currentMerge.takeaways.join(" ");\n mergedVideos.push(currentMerge);\n return mergedVideos;\n};\n\n// Apply merging logic\nconst finalVideos = Object.values(groupedVideos)\n .map(mergeTimeRanges)\n .flat()\n .map(video => {\n // Update link with new start time\n const correctedLink = video.link.replace(/(\?|&)t=\d+s?/, ‘’) + &t=${video.startSeconds}s;\n\n return {\n json: {\n videoTitle: video.videoTitle,\n link: correctedLink,\n timeStamp: video.timeStamp,\n takeaways: video.takeaways\n }\n };\n });\n\n// Return the final structured output\nreturn finalVideos;\n”
},
“type”: “n8n-nodes-base.code”,
“typeVersion”: 2,
“position”: [
-1900,
340
],
“id”: “b8359ad8-562a-435a-95c3-bc60f4c6e143”,
“name”: “Link-inator”
},
{
“parameters”: {
“fieldsToAggregate”: {
“fieldToAggregate”: [
{
“fieldToAggregate”: “takeaways”
}
]
},
“options”: {}
},
“type”: “n8n-nodes-base.aggregate”,
“typeVersion”: 1,
“position”: [
-1700,
340
],
“id”: “e85cea97-c2e1-4e77-9be6-d70c50a3721d”,
“name”: “Aggregate”,
“alwaysOutputData”: false
},
{
“parameters”: {
“jsCode”: “// Get all input items\nconst items = $input.all();\n\n// Generate formatted text array\nconst formattedText = items.map(item => {\n return \"${item.json.videoTitle}\" (${item.json.link}) ${item.json.timeStamp};\n}).join(‘\n\n’);\n\n// Extract takeaways into separate array\nconst takeaways = items.map(item => item.json.takeaways);\n\n// Return both formatted text and takeaways as output\nreturn [{ \n json: { \n formattedText,\n takeaways \n } \n}];”
},
“type”: “n8n-nodes-base.code”,
“typeVersion”: 2,
“position”: [
-1800,
340
],
“id”: “4d2e758d-2180-49bb-82b8-b521f54a75a7”,
“name”: “Neat-inator”
},
{
“parameters”: {
“updates”: [
“message”
],
“additionalFields”: {}
},
“type”: “n8n-nodes-base.telegramTrigger”,
“typeVersion”: 1.1,
“position”: [
-2960,
360
],
“id”: “2a46f371-47bc-4a8c-aac4-1719dd787411”,
“name”: “Telegram Trigger”,
“webhookId”: “06e60453-4ba8-4e51-8666-783b5f4fbeee”,
“credentials”: {
“telegramApi”: {
“id”: “HoA9ebkoEpueVARl”,
“name”: “Tester”
}
}
},
{
“parameters”: {
“rules”: {
“values”: [
{
“conditions”: {
“options”: {
“caseSensitive”: true,
“leftValue”: “”,
“typeValidation”: “strict”,
“version”: 2
},
“conditions”: [
{
“id”: “fb66da42-851e-49f8-90eb-a2ec60eb17f5”,
“leftValue”: “={{ $json.message.text }}”,
“rightValue”: “”,
“operator”: {
“type”: “string”,
“operation”: “exists”,
“singleValue”: true
}
}
],
“combinator”: “and”
},
“renameOutput”: true,
“outputKey”: “Text”
},
{
“conditions”: {
“options”: {
“caseSensitive”: true,
“leftValue”: “”,
“typeValidation”: “strict”,
“version”: 2
},
“conditions”: [
{
“leftValue”: “={{ $json.message.text }}”,
“rightValue”: “”,
“operator”: {
“type”: “string”,
“operation”: “notExists”,
“singleValue”: true
}
}
],
“combinator”: “and”
},
“renameOutput”: true,
“outputKey”: “Audio”
}
]
},
“options”: {}
},
“type”: “n8n-nodes-base.switch”,
“typeVersion”: 3.2,
“position”: [
-2720,
360
],
“id”: “c546adc9-2a73-41bc-b7e6-4b580f3e2f30”,
“name”: “Switch”
},
{
“parameters”: {
“mode”: “retrieve-as-tool”,
“toolName”: “Pinecone”,
“toolDescription”: “Retrieve data from transcripts for satisfying user’s query”,
“pineconeIndex”: {
“__rl”: true,
“value”: “zhang-db”,
“mode”: “list”,
“cachedResultName”: “zhang-db”
},
“topK”: 8,
“options”: {
“pineconeNamespace”: “JZ”
}
},
“type”: “@n8n/n8n-nodes-langchain.vectorStorePinecone”,
“typeVersion”: 1,
“position”: [
-2200,
700
],
“id”: “34858697-a65c-4970-a0b7-3e8cd374ce18”,
“name”: “Javan Zhang”,
“credentials”: {
“pineconeApi”: {
“id”: “77DlESSZEOkmmJN9”,
“name”: “PineconeApi account”
}
}
},
{
“parameters”: {
“chatId”: “={{ $(‘Telegram Trigger’).item.json.message.chat.id }}”,
“text”: “a”,
“additionalFields”: {}
},
“type”: “n8n-nodes-base.telegram”,
“typeVersion”: 1.2,
“position”: [
-1100,
340
],
“id”: “9e30e0e5-3312-4652-be55-2327b0e9d4e9”,
“name”: “Telegram”,
“webhookId”: “835d781b-175f-4004-93e8-768bb21f7af6”,
“credentials”: {
“telegramApi”: {
“id”: “HoA9ebkoEpueVARl”,
“name”: “Tester”
}
}
},
{
“parameters”: {
“assignments”: {
“assignments”: [
{
“id”: “6c918935-d280-43d5-91fa-b7b2b495d5c0”,
“name”: “text”,
“value”: “={{ $json.text }}{{ $json.body.message.toolCallList[0].function.arguments.Query }}{{ $json.chatInput }}{{ $(‘Telegram Trigger’).item.json.message.text }}”,
“type”: “string”
}
]
},
“options”: {}
},
“type”: “n8n-nodes-base.set”,
“typeVersion”: 3.4,
“position”: [
-2460,
340
],
“id”: “2e8e4c94-08f6-44f7-8f9b-1a8a6df1727a”,
“name”: “Body Text”
}
],
“connections”: {
“Embeddings OpenAI2”: {
“ai_embedding”: [
[
{
“node”: “Javan Zhang”,
“type”: “ai_embedding”,
“index”: 0
}
]
]
},
“OpenAI Chat Model1”: {
“ai_languageModel”: [
[
{
“node”: “Deepseeked De-Jsoned”,
“type”: “ai_languageModel”,
“index”: 0
}
]
]
},
“OpenAI1”: {
“main”: [
[
{
“node”: “Telegram”,
“type”: “main”,
“index”: 0
}
]
]
},
“Structured Output Parser”: {
“ai_outputParser”: [
[
{
“node”: “Deepseeked De-Jsoned”,
“type”: “ai_outputParser”,
“index”: 0
}
]
]
},
“Deepseeked De-Jsoned”: {
“main”: [
[
{
“node”: “Link-inator”,
“type”: “main”,
“index”: 0
}
]
]
},
“Link-inator”: {
“main”: [
[
{
“node”: “Neat-inator”,
“type”: “main”,
“index”: 0
}
]
]
},
“Aggregate”: {
“main”: [
[
{
“node”: “OpenAI1”,
“type”: “main”,
“index”: 0
}
]
]
},
“Neat-inator”: {
“main”: [
[
{
“node”: “Aggregate”,
“type”: “main”,
“index”: 0
}
]
]
},
“Telegram Trigger”: {
“main”: [
[
{
“node”: “Switch”,
“type”: “main”,
“index”: 0
}
]
]
},
“Switch”: {
“main”: [
[
{
“node”: “Body Text”,
“type”: “main”,
“index”: 0
}
],
]
},
“Javan Zhang”: {
“ai_tool”: [
[
{
“node”: “Deepseeked De-Jsoned”,
“type”: “ai_tool”,
“index”: 0
}
]
]
},
“Body Text”: {
“main”: [
[
{
“node”: “Deepseeked De-Jsoned”,
“type”: “main”,
“index”: 0
}
]
]
}
},
“pinData”: {},
“meta”: {
“templateCredsSetupCompleted”: true,
“instanceId”: “9d2ce770f045bd348f1244940ac553bce2bc066863672f7c4812dd9dbcc020fd”
}
}
Information on your n8n setup
- n8n version: 1.77.3
- Database (default: SQLite): Default (I assume)
- n8n EXECUTIONS_PROCESS setting (default: own, main): I am not sure where to get that info.
- Running n8n via (Docker, npm, n8n cloud, desktop app): Cloud
- Operating system: Windows 11
If my reply answers your question, please remember to mark it as a solution.