This workflow follows the HTTP Request → OpenAI 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": "Agent Prompt Runner",
"nodes": [
{
"parameters": {
"httpMethod": "POST",
"path": "run-agent-prompt",
"responseMode": "responseNode",
"options": {}
},
"id": "webhook-trigger",
"name": "Webhook",
"type": "n8n-nodes-base.webhook",
"typeVersion": 2,
"position": [
250,
300
]
},
{
"parameters": {
"method": "GET",
"url": "=https://levell-io.vercel.app/api/prompts/{{ $json.body.prompt_id }}",
"options": {}
},
"id": "fetch-prompt",
"name": "Fetch Prompt from API",
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
450,
300
]
},
{
"parameters": {
"method": "GET",
"url": "=https://levell-io.vercel.app/api/transcripts/{{ $('Webhook').first().json.body.transcript_id }}",
"options": {}
},
"id": "fetch-transcript",
"name": "Fetch Transcript",
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
650,
300
]
},
{
"parameters": {
"jsCode": "const webhookData = $('Webhook').first().json.body;\nconst promptResponse = $('Fetch Prompt from API').first().json;\nconst transcript = $input.first().json;\n\n// Access the nested prompt object from API response\nconst prompt = promptResponse.prompt;\n\n// Use test transcript if in test mode, otherwise use fetched transcript\nconst transcriptContent = webhookData.test_mode && webhookData.test_transcript \n ? webhookData.test_transcript \n : (transcript.content || transcript.transcript || '');\n\n// Get system_prompt: webhook override > prompt.system_prompt > prompt.prompt_content\nconst systemPrompt = webhookData.system_prompt \n || prompt.system_prompt \n || prompt.prompt_content \n || '';\n\n// Get user_prompt_template: webhook override > prompt.user_prompt_template\nconst userPromptTemplate = webhookData.user_prompt_template \n || prompt.user_prompt_template \n || '';\n\n// Get temperature: webhook override > prompt.temperature > default 0.3\nconst temperature = webhookData.temperature ?? prompt.temperature ?? 0.3;\n\n// Build the user message by replacing {{transcript}} and other placeholders\nlet userMessage = userPromptTemplate\n ? userPromptTemplate\n .replace(/\\{\\{transcript\\}\\}/gi, transcriptContent)\n .replace(/\\{\\{rep_transcript_name\\}\\}/gi, webhookData.test_transcript_name || '')\n : `Analyze this sales call transcript:\\n\\n${transcriptContent}`;\n\n// IMPORTANT: Preserve transcript_id from webhook body\nconst transcriptId = webhookData.transcript_id;\n\nreturn {\n prompt_id: prompt.id,\n prompt_version: prompt.version || 1,\n agent_type: prompt.agent_type,\n system_prompt: systemPrompt,\n user_prompt_template: userPromptTemplate,\n user_message: userMessage,\n temperature: temperature,\n transcript: transcriptContent,\n transcript_id: transcriptId,\n user_id: webhookData.user_id || null,\n test_mode: webhookData.test_mode || false\n};"
},
"id": "prepare-prompt",
"name": "Prepare Prompt Data",
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
850,
300
]
},
{
"parameters": {
"modelId": {
"__rl": true,
"value": "gpt-4.1",
"mode": "list",
"cachedResultName": "GPT-4.1"
},
"messages": {
"values": [
{
"content": "={{ $json.system_prompt }}",
"role": "system"
},
{
"content": "={{ $json.user_message }}"
}
]
},
"options": {
"maxTokens": 4096,
"temperature": "={{ $json.temperature }}",
"responseFormat": "json_object"
}
},
"id": "openai-call",
"name": "OpenAI Chat",
"type": "@n8n/n8n-nodes-langchain.openAi",
"typeVersion": 1.8,
"position": [
1050,
300
],
"credentials": {
"openAiApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"jsCode": "const promptData = $('Prepare Prompt Data').first().json;\nconst aiResponse = $input.first().json;\n\n// Parse the AI response\nlet outputData = {};\ntry {\n outputData = JSON.parse(aiResponse.message?.content || '{}');\n} catch (e) {\n outputData = { raw_output: aiResponse.message?.content };\n}\n\nconst usage = aiResponse.usage || {};\n\n// Get the actual model used from the response\nconst modelUsed = aiResponse.model || 'gpt-4.1';\n\n// Calculate cost based on model (rates per 1M tokens)\nconst inputTokens = usage.prompt_tokens || 0;\nconst outputTokens = usage.completion_tokens || 0;\n\nlet inputRate = 2.5;\nlet outputRate = 10.0;\n\nif (modelUsed.includes('gpt-4.1') || modelUsed.includes('gpt-4-turbo')) {\n inputRate = 2.0;\n outputRate = 8.0;\n} else if (modelUsed.includes('gpt-4o-mini')) {\n inputRate = 0.15;\n outputRate = 0.60;\n} else if (modelUsed.includes('gpt-4o')) {\n inputRate = 2.5;\n outputRate = 10.0;\n}\n\nconst inputCost = (inputTokens * inputRate) / 1000000;\nconst outputCost = (outputTokens * outputRate) / 1000000;\nconst totalCost = inputCost + outputCost;\n\nreturn {\n // Prompt identification\n prompt_id: promptData.prompt_id,\n prompt_version: promptData.prompt_version,\n agent_type: promptData.agent_type,\n \n // Prompt content (for saving to DB)\n system_prompt: promptData.system_prompt,\n user_message: promptData.user_message,\n temperature: promptData.temperature,\n \n // Run metadata\n run_type: 'n8n',\n is_test_run: promptData.test_mode,\n transcript_id: promptData.transcript_id,\n user_id: promptData.user_id,\n \n // AI Response\n output: aiResponse.message?.content || '',\n output_data: outputData,\n \n // Token usage\n input_tokens: inputTokens,\n output_tokens: outputTokens,\n prompt_tokens: inputTokens,\n completion_tokens: outputTokens,\n total_tokens: inputTokens + outputTokens,\n \n // Cost calculation\n cost_usd: totalCost,\n total_cost: totalCost,\n \n // Model info\n model: modelUsed,\n status: 'completed'\n};"
},
"id": "prepare-run-data",
"name": "Prepare Run Data",
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
1250,
300
]
},
{
"parameters": {
"method": "POST",
"url": "https://levell-io.vercel.app/api/agent-runs",
"sendBody": true,
"specifyBody": "json",
"jsonBody": "={{ JSON.stringify({\n agent_type: $json.agent_type,\n prompt_id: $json.prompt_id,\n prompt_sent: $json.system_prompt,\n system_prompt: $json.system_prompt,\n user_message: $json.user_message,\n output: $json.output,\n model: $json.model,\n prompt_tokens: $json.prompt_tokens,\n completion_tokens: $json.completion_tokens,\n transcript_id: $json.transcript_id,\n user_id: $json.user_id,\n context_type: $json.run_type,\n status: $json.status,\n metadata: {\n prompt_version: $json.prompt_version,\n is_test_run: $json.is_test_run,\n cost_usd: $json.cost_usd,\n temperature: $json.temperature,\n input_tokens: $json.input_tokens,\n output_tokens: $json.output_tokens,\n total_tokens: $json.total_tokens\n }\n}) }}",
"options": {}
},
"id": "save-run",
"name": "Save Run to API",
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4.2,
"position": [
1450,
300
]
},
{
"parameters": {
"respondWith": "json",
"responseBody": "={{ JSON.stringify({\n success: true,\n run_id: $json.run?.id || $json.id,\n agent_type: $('Prepare Run Data').first().json.agent_type,\n prompt_version: $('Prepare Run Data').first().json.prompt_version,\n message: 'Agent run completed',\n transcript_id: $('Prepare Run Data').first().json.transcript_id,\n temperature: $('Prepare Run Data').first().json.temperature,\n token_usage: {\n input: $('Prepare Run Data').first().json.input_tokens,\n output: $('Prepare Run Data').first().json.output_tokens,\n total: $('Prepare Run Data').first().json.total_tokens\n },\n cost_usd: $('Prepare Run Data').first().json.cost_usd,\n model: $('Prepare Run Data').first().json.model\n}) }}"
},
"id": "respond",
"name": "Respond to Webhook",
"type": "n8n-nodes-base.respondToWebhook",
"typeVersion": 1.1,
"position": [
1650,
300
]
}
],
"connections": {
"Webhook": {
"main": [
[
{
"node": "Fetch Prompt from API",
"type": "main",
"index": 0
}
]
]
},
"Fetch Prompt from API": {
"main": [
[
{
"node": "Fetch Transcript",
"type": "main",
"index": 0
}
]
]
},
"Fetch Transcript": {
"main": [
[
{
"node": "Prepare Prompt Data",
"type": "main",
"index": 0
}
]
]
},
"Prepare Prompt Data": {
"main": [
[
{
"node": "OpenAI Chat",
"type": "main",
"index": 0
}
]
]
},
"OpenAI Chat": {
"main": [
[
{
"node": "Prepare Run Data",
"type": "main",
"index": 0
}
]
]
},
"Prepare Run Data": {
"main": [
[
{
"node": "Save Run to API",
"type": "main",
"index": 0
}
]
]
},
"Save Run to API": {
"main": [
[
{
"node": "Respond to Webhook",
"type": "main",
"index": 0
}
]
]
}
},
"settings": {
"executionOrder": "v1"
},
"meta": {
"templateCredsSetupCompleted": true
}
}
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.
openAiApi
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
Agent Prompt Runner. Uses httpRequest, openAi. Webhook trigger; 8 nodes.
Source: https://github.com/rajpalom13/levell.io/blob/6a1ddb7bafe9eeff8fcf0922dfc059531178a90e/n8n-workflows/agent-prompt-runner.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.
CLINICAINTEGRAL_secretary. Uses postgres, mcpClientTool, googleDriveTool, toolWorkflow. Webhook trigger; 89 nodes.
Postgres, Mcp Client Tool, Google Drive Tool +14
Remi 1.1. Uses lmChatOpenAi, memoryPostgresChat, openAi, postgres. Webhook trigger; 89 nodes.
OpenAI Chat, Memory Postgres Chat, OpenAI +7
This n8n workflow orchestrates a powerful suite of AI Agents and automations to manage and optimize various aspects of an e-commerce operation, particularly for platforms like Shopify. It leverages La
Google Sheets, HTTP Request, Slack +10
My workflow 7. Uses openAi, redis, httpRequest, agent. Webhook trigger; 77 nodes.
OpenAI, Redis, HTTP Request +8
What if AI didn't just write content—but actually thought about how to write it? This n8n workflow revolutionizes content creation by deploying multiple specialized AI agents that handle every aspect
Tool Http Request, Anthropic Chat, Airtable +7