This workflow follows the Chainllm → Informationextractor recipe pattern — see all workflows that pair these two integrations.
The workflow JSON
Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →
{
"name": "Chain LLM Example: Customer Email Triage (Groq)",
"nodes": [
{
"parameters": {
"httpMethod": "POST",
"path": "customer-email-triage",
"responseMode": "responseNode",
"options": {}
},
"id": "a1b2c3d4-0001-0001-0001-000000000001",
"name": "Webhook",
"type": "n8n-nodes-base.webhook",
"typeVersion": 2,
"position": [
240,
300
]
},
{
"parameters": {
"schemaType": "manual",
"inputSchema": "{\n \"type\": \"object\",\n \"properties\": {\n \"customer_name\": { \"type\": \"string\" },\n \"product_mentioned\": { \"type\": \"string\" },\n \"language\": { \"type\": \"string\", \"enum\": [\"en\", \"es\", \"pt\", \"fr\", \"de\", \"other\"] },\n \"sentiment\": { \"type\": \"string\", \"enum\": [\"positive\", \"neutral\", \"negative\"] },\n \"urgency_score\": { \"type\": \"number\", \"minimum\": 0, \"maximum\": 10 },\n \"contains_pii\": { \"type\": \"boolean\" }\n },\n \"required\": [\"customer_name\", \"sentiment\", \"urgency_score\"]\n}",
"text": "={{ $json.body.email_content }}",
"options": {
"systemPromptTemplate": "You extract structured data from customer emails. Extract only what is explicitly stated. Do not infer or guess. If a field is not present in the email, omit it. Detect language from the actual email content. PII includes phone numbers, addresses, full credit card numbers, or government IDs."
}
},
"id": "a1b2c3d4-0001-0001-0001-000000000002",
"name": "Stage 1: Information Extractor",
"type": "@n8n/n8n-nodes-langchain.informationExtractor",
"typeVersion": 1,
"position": [
480,
300
]
},
{
"parameters": {
"model": "llama-3.3-70b-versatile",
"options": {
"temperature": 0,
"maxTokensToSample": 500
}
},
"id": "a1b2c3d4-0001-0001-0001-000000000003",
"name": "Groq Chat (Stage 1)",
"type": "@n8n/n8n-nodes-langchain.lmChatGroq",
"typeVersion": 1,
"position": [
480,
480
],
"credentials": {
"groqApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"mode": "runOnceForEachItem",
"jsCode": "// Validate Stage 1 output before passing to Stage 2\nconst data = $input.item.json.output;\n\nif (!data || !data.customer_name || !data.sentiment || data.urgency_score == null) {\n throw new Error(`Stage 1 missing required fields: ${JSON.stringify(data)}`);\n}\n\nif (data.urgency_score < 0 || data.urgency_score > 10) {\n throw new Error(`urgency_score out of range: ${data.urgency_score}`);\n}\n\nreturn { json: { extracted: data, original_email: $('Webhook').item.json.body.email_content } };"
},
"id": "a1b2c3d4-0001-0001-0001-000000000004",
"name": "Validate Stage 1",
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
720,
300
]
},
{
"parameters": {
"promptType": "define",
"text": "=Given this extracted customer data:\n\n{{ JSON.stringify($json.extracted, null, 2) }}\n\nClassify into one of: technical_issue, billing_question, cancellation_risk, upsell_opportunity, compliment, other.\n\nThen identify the single most important next action (one short sentence).\n\nReturn JSON with keys: category, next_action, reasoning. No prose, no markdown.",
"options": {
"systemMessage": "You are a customer support triage analyst. You return only valid JSON. You do not invent categories outside the allowed list."
}
},
"id": "a1b2c3d4-0001-0001-0001-000000000005",
"name": "Stage 2: Classify + Action",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"typeVersion": 1.4,
"position": [
960,
300
]
},
{
"parameters": {
"model": "llama-3.3-70b-versatile",
"options": {
"temperature": 0.2,
"maxTokensToSample": 400,
"responseFormat": "json_object"
}
},
"id": "a1b2c3d4-0001-0001-0001-000000000006",
"name": "Groq Chat (Stage 2)",
"type": "@n8n/n8n-nodes-langchain.lmChatGroq",
"typeVersion": 1,
"position": [
960,
480
],
"credentials": {
"groqApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"mode": "runOnceForEachItem",
"jsCode": "// Stage 3: Deterministic scoring (NO LLM \u2014 math goes in Code, not in LLM)\nconst stage1 = $('Validate Stage 1').item.json.extracted;\nconst stage2 = JSON.parse($input.item.json.text);\n\n// Composite priority score: 0.5 * urgency + 0.3 * sentiment_weight + 0.2 * category_weight\nconst sentimentWeight = stage1.sentiment === 'negative' ? 10 : stage1.sentiment === 'neutral' ? 5 : 1;\nconst categoryWeight = {\n cancellation_risk: 10,\n technical_issue: 8,\n billing_question: 6,\n upsell_opportunity: 4,\n compliment: 1,\n other: 3\n}[stage2.category] ?? 3;\n\nconst priority = (0.5 * stage1.urgency_score) + (0.3 * sentimentWeight) + (0.2 * categoryWeight);\nconst routeTo = priority >= 7 ? 'urgent_queue' : priority >= 4 ? 'standard_queue' : 'low_queue';\n\nreturn {\n json: {\n customer: stage1.customer_name,\n language: stage1.language,\n contains_pii: stage1.contains_pii ?? false,\n sentiment: stage1.sentiment,\n category: stage2.category,\n next_action: stage2.next_action,\n priority_score: Number(priority.toFixed(2)),\n route_to: routeTo,\n reasoning: stage2.reasoning,\n processed_at: new Date().toISOString()\n }\n};"
},
"id": "a1b2c3d4-0001-0001-0001-000000000007",
"name": "Stage 3: Score + Route (Code)",
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
1200,
300
]
},
{
"parameters": {
"respondWith": "json",
"responseBody": "={{ JSON.stringify($json) }}"
},
"id": "a1b2c3d4-0001-0001-0001-000000000008",
"name": "Respond",
"type": "n8n-nodes-base.respondToWebhook",
"typeVersion": 1,
"position": [
1440,
300
]
}
],
"connections": {
"Webhook": {
"main": [
[
{
"node": "Stage 1: Information Extractor",
"type": "main",
"index": 0
}
]
]
},
"Stage 1: Information Extractor": {
"main": [
[
{
"node": "Validate Stage 1",
"type": "main",
"index": 0
}
]
]
},
"Groq Chat (Stage 1)": {
"ai_languageModel": [
[
{
"node": "Stage 1: Information Extractor",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Validate Stage 1": {
"main": [
[
{
"node": "Stage 2: Classify + Action",
"type": "main",
"index": 0
}
]
]
},
"Stage 2: Classify + Action": {
"main": [
[
{
"node": "Stage 3: Score + Route (Code)",
"type": "main",
"index": 0
}
]
]
},
"Groq Chat (Stage 2)": {
"ai_languageModel": [
[
{
"node": "Stage 2: Classify + Action",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Stage 3: Score + Route (Code)": {
"main": [
[
{
"node": "Respond",
"type": "main",
"index": 0
}
]
]
}
},
"settings": {
"executionOrder": "v1",
"saveDataErrorExecution": "all",
"saveDataSuccessExecution": "all"
},
"tags": [
{
"name": "chain-llm-pattern"
},
{
"name": "example"
}
]
}
Credentials you'll need
Each integration node will prompt for credentials when you import. We strip credential IDs before publishing — you'll add your own.
groqApi
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
Chain LLM Example: Customer Email Triage (Groq). Uses informationExtractor, lmChatGroq, chainLlm. Webhook trigger; 8 nodes.
Source: https://github.com/masteranime/n8n-claude-skills/blob/4cb50176ad2fedf010bbfb5a5d1f6b049a90348c/examples/groq-chain-example.json — original creator credit. Request a take-down →
Related workflows
Workflows that share integrations, category, or trigger type with this one. All free to copy and import.
Create and publish Instagram carousels using OpenAI gpt-image-1 and AI caption. Uses chainLlm, outputParserItemList, lmChatOpenAi, splitInBatches. Scheduled trigger; 32 nodes.
Chain Llm, Output Parser Item List, OpenAI Chat +3
Vacancy Launch: Calendar, ClickUp & AI LinkedIn Post. Uses googleCalendar, stickyNote, clickUp, linkedIn. Webhook trigger; 18 nodes.
Google Calendar, ClickUp, LinkedIn +1
Sia — Portal de Expansão de Terrenos (Principal). Uses chainLlm, lmChatGroq, googleSheets. Webhook trigger; 12 nodes.
Chain Llm, Groq Chat, Google Sheets
GiveWP Donations to Beacon. Uses httpRequest, stopAndError. Webhook trigger; 43 nodes.
HTTP Request, Stop And Error
Use this workflow to book, cancel, or reschedule appointments using Vapi and Google Calendar
Google Calendar