Dr. Firas This workflow corresponds to n8n.io template #17597 — we link there as the canonical source.
This workflow follows the Chainllm → Chat Trigger 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 →
{
"id": "yYF8xR5q4y9D1DuL",
"meta": {
"builderVariant": "mcp",
"aiBuilderAssisted": true
},
"name": "Generate sound effects from chat using OpenAI and ElevenLabs",
"tags": [],
"nodes": [
{
"id": "eb94f891-c347-43d7-80d3-ac432498de73",
"name": "When Chat Message Received",
"type": "@n8n/n8n-nodes-langchain.chatTrigger",
"position": [
-544,
224
],
"parameters": {
"options": {
"responseMode": "lastNode"
}
},
"typeVersion": 1.4
},
{
"id": "3571324e-8bfb-4ae4-a577-58dbb6f90b37",
"name": "Compose Sound Effect Prompt",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"onError": "continueErrorOutput",
"position": [
-128,
224
],
"parameters": {
"text": "={{ $json.chatInput }}",
"batching": {},
"messages": {
"messageValues": [
{
"message": "You are an expert sound-effects director for the ElevenLabs sound-generation API. Turn the user's casual description into ONE richly detailed, literal sound-effects prompt, then choose sensible settings.\n\nWrite the prompt so it describes the sound literally: the source, material, action, environment, and how it evolves over time. Prefer concrete audio detail (for example, \"a heavy oak door slowly creaking open, hinges groaning, faint echo in a stone hall\") over vague labels (for example, \"door sound\"). Keep it to one continuous sound event unless the user clearly asks for a sequence. Do not add music or narration unless asked.\n\nPick duration_seconds between 0.5 and 30 (0.5 to 2 for short impacts like clicks or thuds, 8 to 20 for ambiences or drones). Pick prompt_influence between 0 and 1 (about 0.3 for organic or natural sounds, 0.6 to 0.8 for precise, mechanical, or branded sounds).\n\nReturn ONLY a JSON object and nothing else. No markdown, no code fences, no commentary. Use exactly these keys:\n{\"sfx_prompt\": \"a literal detailed description\", \"duration_seconds\": 6, \"prompt_influence\": 0.3}"
}
]
},
"promptType": "define"
},
"typeVersion": 1.9
},
{
"id": "41faa985-9c3f-4f70-8429-c73fd9674fc2",
"name": "OpenAI Sound Director Model",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
-128,
400
],
"parameters": {
"model": {
"__rl": true,
"mode": "list",
"value": "gpt-5-mini"
},
"options": {
"temperature": 0.2
},
"builtInTools": {}
},
"credentials": {
"openAiApi": {
"name": "<your credential>"
}
},
"typeVersion": 1.3
},
{
"id": "1e825253-3ccd-4243-967a-7891fc7bb8d3",
"name": "Clamp Sound Settings",
"type": "n8n-nodes-base.code",
"position": [
208,
144
],
"parameters": {
"mode": "runOnceForEachItem",
"jsCode": "// Parse the sound director's JSON reply, then normalize and clamp it.\n// Robust by design: any parse failure falls back to the user's own words,\n// so an empty, garbled, or non-JSON reply still yields a valid request.\n\nconst rawMessage = ($('When Chat Message Received').item.json.chatInput || '').toString().trim();\n\n// Basic LLM Chain returns the model text; be liberal about where it sits.\nconst text = ($json.text ?? $json.response?.text ?? $json.output ?? $json.data ?? '').toString();\n\nlet parsed = {};\ntry {\n // Drop any markdown fences, then grab the first {...} block if prose surrounds it.\n const cleaned = text.replace(/```(?:json)?/gi, '').trim();\n const match = cleaned.match(/\\{[\\s\\S]*\\}/);\n parsed = JSON.parse(match ? match[0] : cleaned);\n} catch (e) {\n parsed = {};\n}\n\nconst clamp = (n, lo, hi) => Math.min(hi, Math.max(lo, n));\nconst toNum = (v) => {\n const n = typeof v === 'number' ? v : parseFloat(v);\n return Number.isFinite(n) ? n : null;\n};\n\n// Prompt: prefer the director's rewrite, fall back to the raw words, then a safe default.\nlet sfxPrompt = (parsed && typeof parsed.sfx_prompt === 'string') ? parsed.sfx_prompt.trim() : '';\nif (!sfxPrompt) sfxPrompt = rawMessage;\nif (!sfxPrompt) sfxPrompt = 'a soft, short ambient tone';\n\n// Duration: clamp to the ElevenLabs 0.5-30s window; default 3s if missing.\nlet duration = toNum(parsed ? parsed.duration_seconds : null);\nduration = duration === null ? 3 : Math.round(clamp(duration, 0.5, 30) * 10) / 10;\n\n// Prompt influence: clamp to 0-1; default 0.3 (the ElevenLabs default).\nlet influence = toNum(parsed ? parsed.prompt_influence : null);\ninfluence = influence === null ? 0.3 : clamp(influence, 0, 1);\n\nreturn { json: { sfx_prompt: sfxPrompt, duration_seconds: duration, prompt_influence: influence } };"
},
"typeVersion": 2
},
{
"id": "d4bfacc8-dedc-43c9-896d-f28ee52d7ad3",
"name": "Generate Sound With ElevenLabs",
"type": "n8n-nodes-base.httpRequest",
"onError": "continueErrorOutput",
"maxTries": 3,
"position": [
400,
224
],
"parameters": {
"url": "https://api.elevenlabs.io/v1/sound-generation",
"method": "POST",
"options": {
"timeout": 120000,
"response": {
"response": {
"responseFormat": "file"
}
}
},
"jsonBody": "={{ JSON.stringify({ text: $json.sfx_prompt, duration_seconds: $json.duration_seconds, prompt_influence: $json.prompt_influence }) }}",
"sendBody": true,
"sendQuery": true,
"specifyBody": "json",
"authentication": "genericCredentialType",
"genericAuthType": "httpHeaderAuth",
"queryParameters": {
"parameters": [
{
"name": "output_format",
"value": "mp3_44100_128"
}
]
}
},
"credentials": {
"httpHeaderAuth": {
"name": "<your credential>"
}
},
"retryOnFail": true,
"typeVersion": 4.4,
"waitBetweenTries": 5000
},
{
"id": "df10b8f1-88b8-4ac9-a9e2-41f071b8cf11",
"name": "Reply In Chat With Sound",
"type": "n8n-nodes-base.set",
"position": [
608,
80
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "64f5195f-b64c-433e-999d-8a2747ff43a5",
"name": "output",
"type": "string",
"value": "=\ud83d\udd0a Your sound effect is ready, listen to it above.\n\n**Prompt used:** {{ $('Clamp Sound Settings').item.json.sfx_prompt }}\n**Duration:** {{ $('Clamp Sound Settings').item.json.duration_seconds }}s \u00b7 **Prompt influence:** {{ $('Clamp Sound Settings').item.json.prompt_influence }}"
}
]
},
"includeOtherFields": true
},
"typeVersion": 3.4
},
{
"id": "6bbc04fd-28cc-4d67-b214-c878c2eedd0a",
"name": "Report Sound Failure",
"type": "n8n-nodes-base.set",
"position": [
608,
288
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "67d98c22-f143-40cd-ba1e-a55567153163",
"name": "output",
"type": "string",
"value": "\u26a0\ufe0f I couldn't generate that sound effect just now. The sound service may be busy, or the description was unclear. Please try again with a clear description like \"a heavy wooden door creaking open\" or \"distant thunder rumbling\"."
}
]
}
},
"typeVersion": 3.4
},
{
"id": "bdb47582-df10-4bac-8370-184f29d6a7d5",
"name": "Sticky Note 3737cb97",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1488,
-224
],
"parameters": {
"width": 672,
"height": 1548,
"content": "# Generate sound effects from chat using OpenAI and ElevenLabs\n# \ud83d\udce5 [Open full documentation on Notion](https://automatisation.notion.site/Course-Generate-sound-effects-from-chat-using-OpenAI-and-ElevenLabs-39f3d6550fd9818aa664cfde59a5d1cd)\n\n## How it works\n\n1. The user describes a sound in plain words in the n8n chat.\n2. An OpenAI model acts as a sound-effects director and rewrites the request into a literal, richly detailed ElevenLabs prompt, then picks a duration and a prompt influence.\n3. A Code node parses that JSON and clamps the settings to safe ranges, then calls the ElevenLabs sound-generation API.\n4. The generated MP3 is attached directly to the chat reply (no external storage needed) along with the prompt and settings used.\n\n## Setup\n\n- [ ] Select your \"OpenAi account\" credential on OpenAI Sound Director Model.\n- [ ] Create a Header Auth credential named \"ElevenLabs API\" with header name xi-api-key and your ElevenLabs API key, then select it on Generate Sound With ElevenLabs.\n\n## Requirements\n\n- An OpenAI API key (OpenAi account credential).\n- An ElevenLabs API key (Header Auth credential, header name xi-api-key).\n- n8n Cloud or self-hosted with the built-in Chat Trigger enabled.\n\n## Customization\n\nSwap the chat trigger for a form or webhook, tune the duration and prompt-influence guidance in the director's system prompt, or change the OpenAI model used for sound direction.\n\nNeed help customizing?\nContact me for consulting and support : [Linkedin](https://www.linkedin.com/in/doctor-firass/)\n\n# MY NEW YOUTUBE CHANNEL\n\ud83d\udc49 [Subscribe to my new YouTube channel](https://www.youtube.com/@DrFiras_AI). Here I'll share videos and Shorts with practical tutorials and FREE templates for n8n.\n\n[](https://www.youtube.com/@DrFiras_AI)"
},
"typeVersion": 1
},
{
"id": "2eafdbef-0b55-41da-8d36-9b7733d7a60b",
"name": "Sticky Note e0973466",
"type": "n8n-nodes-base.stickyNote",
"position": [
-176,
-240
],
"parameters": {
"color": 7,
"width": 428,
"height": 300,
"content": "## Describe and direct the sound\n\nTakes the chat message and turns it into a clean, literal sound prompt with parsed, clamped settings."
},
"typeVersion": 1
},
{
"id": "d6f37dcb-959b-477e-9690-e044886bb28d",
"name": "Sticky Note adf541be",
"type": "n8n-nodes-base.stickyNote",
"position": [
272,
-240
],
"parameters": {
"color": 7,
"width": 444,
"height": 296,
"content": "## Generate the sound effect\n\nBuilds a safely escaped JSON body and calls ElevenLabs, with retries and an error path."
},
"typeVersion": 1
},
{
"id": "8cef16a2-925f-4575-ac41-ea5ad4a867f4",
"name": "Sticky Note e66a09f4",
"type": "n8n-nodes-base.stickyNote",
"position": [
-672,
-240
],
"parameters": {
"color": 7,
"width": 476,
"height": 308,
"content": "## Reply in chat\n\nAttaches the generated MP3 directly to the chat reply, or sends a friendly message on failure. No external storage required."
},
"typeVersion": 1
}
],
"active": true,
"settings": {
"binaryMode": "separate",
"availableInMCP": true,
"executionOrder": "v1"
},
"versionId": "7d57d672-4b4a-48fb-ab73-ea3943397f80",
"nodeGroups": [],
"connections": {
"Clamp Sound Settings": {
"main": [
[
{
"node": "Generate Sound With ElevenLabs",
"type": "main",
"index": 0
}
]
]
},
"When Chat Message Received": {
"main": [
[
{
"node": "Compose Sound Effect Prompt",
"type": "main",
"index": 0
}
]
]
},
"Compose Sound Effect Prompt": {
"main": [
[
{
"node": "Clamp Sound Settings",
"type": "main",
"index": 0
}
],
[
{
"node": "Clamp Sound Settings",
"type": "main",
"index": 0
}
]
]
},
"OpenAI Sound Director Model": {
"ai_languageModel": [
[
{
"node": "Compose Sound Effect Prompt",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Generate Sound With ElevenLabs": {
"main": [
[
{
"node": "Reply In Chat With Sound",
"type": "main",
"index": 0
}
],
[
{
"node": "Report Sound Failure",
"type": "main",
"index": 0
}
]
]
}
}
}
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.
httpHeaderAuthopenAiApi
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
This workflow turns an n8n chat message into a detailed sound-effect prompt using OpenAI, then generates an MP3 sound effect via the ElevenLabs Sound Generation API and replies in chat with the audio and the settings used. Receives a message in n8n Chat describing the sound…
Source: https://n8n.io/workflows/17597/ — 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.
This Chatbot automates the process of discovering job openings and generating tailored job application emails.
Chat Trigger, OpenAI Chat, Mcp Client Tool +12
Job Application PredictLeads & ScrapeGraph AI. Uses chatTrigger, lmChatOpenAi, mcpClientTool, memoryBufferWindow. Chat trigger; 32 nodes.
Chat Trigger, OpenAI Chat, Mcp Client Tool +12
Job Application PredictLeads & ScrapeGraph AI. Uses chatTrigger, lmChatOpenAi, mcpClientTool, memoryBufferWindow. Chat trigger; 32 nodes.
Chat Trigger, OpenAI Chat, Mcp Client Tool +12
This workflow is perfect for developers, researchers, and data scientists benchmarking multiple LLMs with LM Studio. It dynamically fetches active models, tests prompts, and tracks metrics like word c
HTTP Request, Chat Trigger, OpenAI Chat +2
This template provides a high-performance, cost-optimized alternative to standard AI Agents for building RAG (Retrieval-Augmented Generation) chatbots.
Chat Trigger, OpenAI Chat, Chain Llm +5