January 22, 2026

Google Sheets + Google Drive, ad images delivered fast

Lisa Granqvist Partner Workflow Automation Expert

Get a free AI assessment → ⬇️ Use template

Launching new ad creatives shouldn’t feel like a scavenger hunt through spreadsheets, folders, and half-finished prompts. But when every product needs “just a few more variations,” the process turns into constant copy-paste, broken links, and review chaos.

This ad image automation hits performance marketers first. But agency owners and in-house creative ops feel the same drag when they’re trying to ship 20 to 50 variations fast without losing track of what’s approved.

This n8n workflow turns Google Sheets rows into UGC-style ad image prompts, generates the images, stores them in Google Drive, and writes clean links back to your Sheet. You’ll see how it works, what you need, and what to watch out for.

How This Automation Works

The full n8n workflow, from trigger to final output:

n8n Workflow Template: Google Sheets + Google Drive, ad images delivered fast

Click to explore

flowchart LR

    subgraph sg0["When clicking ‘Execute workflow’ Flow"]
        direction LR
        n0@{ icon: "mdi:play-circle", form: "rounded", label: "When clicking ‘Execute workf..", pos: "b", h: 48 }
        n1@{ icon: "mdi:database", form: "rounded", label: "Get row(s) in sheet", pos: "b", h: 48 }
        n2@{ icon: "mdi:robot", form: "rounded", label: "Analyze image", pos: "b", h: 48 }
        n3@{ icon: "mdi:cog", form: "rounded", label: "Download file", pos: "b", h: 48 }
        n4@{ icon: "mdi:robot", form: "rounded", label: "AI Agent", pos: "b", h: 48 }
        n5@{ icon: "mdi:brain", form: "rounded", label: "OpenAI Chat Model", pos: "b", h: 48 }
        n6@{ icon: "mdi:swap-vertical", form: "rounded", label: "Split Out", pos: "b", h: 48 }
        n7@{ icon: "mdi:database", form: "rounded", label: "Append row in sheet", pos: "b", h: 48 }
        n8@{ icon: "mdi:cog", form: "rounded", label: "Wait", pos: "b", h: 48 }
        n9@{ icon: "mdi:swap-horizontal", form: "rounded", label: "If", pos: "b", h: 48 }
        n10["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>Get image status"]
        n11["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>Call Fal.ai API (nannoBanana)"]
        n12@{ icon: "mdi:database", form: "rounded", label: "Update row in sheet", pos: "b", h: 48 }
        n13@{ icon: "mdi:cog", form: "rounded", label: "Upload file", pos: "b", h: 48 }
        n14@{ icon: "mdi:swap-vertical", form: "rounded", label: "Edit Fields", pos: "b", h: 48 }
        n15@{ icon: "mdi:robot", form: "rounded", label: "Structured Output Parser", pos: "b", h: 48 }
        n16["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>HTTP Request"]
        n17["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>Get the image1"]
        n9 --> n17
        n9 --> n8
        n8 --> n10
        n4 --> n6
        n6 --> n7
        n14 --> n11
        n13 --> n12
        n16 --> n13
        n2 --> n4
        n3 --> n2
        n17 --> n16
        n10 --> n9
        n5 -.-> n4
        n7 --> n14
        n1 --> n3
        n15 -.-> n4
        n11 --> n10
        n0 --> n1
    end

    %% Styling
    classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
    classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
    classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
    classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
    classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
    classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
    classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
    classDef disabled stroke-dasharray: 5 5,opacity: 0.5
    class n0 trigger
    class n2,n4,n15 ai
    class n5 aiModel
    class n9 decision
    class n1,n7,n12 database
    class n10,n11,n16,n17 api
    classDef customIcon fill:none,stroke:none
    class n10,n11,n16,n17 customIcon

The Problem: Ad image variations are a tracking nightmare

If you’ve ever tried to scale UGC-style ad variations from a product catalog, you know the ugly part isn’t the “idea.” It’s the coordination. Product info lives in a sheet, the hero image lives in Drive, prompts live in someone’s doc, and outputs end up scattered across chats. Then somebody asks, “Which version did we ship?” and you’re spending an hour reconstructing the trail from memory. Multiply that by a handful of SKUs and a couple campaigns, and you’re suddenly managing a creative factory with duct tape.

The friction compounds. Here’s where things usually snap.

You generate variations, but the prompts and outputs aren’t stored together, so review turns into guesswork.
Drive links break or require access, which means your image tools can’t fetch the input files reliably.
Manual status tracking (“queued,” “rendering,” “done”) gets stale fast, so teams duplicate work.
One missing field in the spreadsheet forces you to pause, fix, rerun, and remember where you left off.

The Solution: Turn each sheet row into ready-to-review ad images

This workflow uses Google Sheets as your control panel and Google Drive as your creative library. It starts by pulling product rows (name, description, constraints, number of variations, aspect ratio, and more). Next, it fetches the product image from Drive, runs a quick image inspection with OpenAI (Vision/Chat), and generates structured prompts for UGC-style scenes using an “orchestrator” step that keeps the output consistent. Those prompts get appended into a separate ad_image sheet so you can review or iterate without touching the generator logic.

When you’re ready to render, the workflow sends each prompt to Fal.ai’s nano-banana model, polls the status until the job completes, downloads the finished image, uploads it to Google Drive, then updates the sheet with a clean output URL. Prompts, statuses, and links stay tied to the same record, so you can actually scale this without losing your mind.

What You Get: Automation vs. Results

What This Workflow Automates

Results You’ll Get

It reads product rows from Google Sheets and standardizes the input fields.
It converts messy Drive share links into direct, fetchable image URLs.
It analyzes the product image and writes structured UGC-style prompts back to Sheets.
It generates image variations via Fal.ai, uploads outputs to Drive, and updates each row with the final link.

Most teams get back about 5 hours a week once variations ramp up.
You stop losing “the good prompt” because prompts and outputs stay connected.
Creative review gets simpler since every output has a stable Drive link.
Fewer reruns, because statuses and failures are visible in the sheet.
You can scale from a few SKUs to dozens without changing the process.

Example: What This Looks Like

Say you’re launching 10 SKUs and you want 5 UGC-style image variations per SKU. Manually, you’re typically doing about 10 minutes per variation between downloading the product image, rewriting prompts, running generation, uploading to Drive, and pasting links into a sheet, which is roughly 8 hours of tedious work. With this workflow, you fill in the row once, trigger the run, and then you’re mostly waiting on rendering. Realistically, you spend about 30 minutes checking prompts and approving outputs instead of babysitting every file.

What You’ll Need

n8n instance (try n8n Cloud free)
Self-hosting option if you prefer (Hostinger works well)
Google Sheets to store inputs, prompts, and status.
Google Drive to fetch inputs and store outputs.
Fal.ai API key (set env var FAL_KEY in n8n).
OpenAI API access (use it for image analysis and prompt generation).

Skill level: Intermediate. You’ll connect OAuth credentials, set one environment variable, and confirm your Sheet columns match the workflow.

Don’t want to set this up yourself? Talk to an automation expert (free 15-minute consultation).

How It Works

A manual run (or scheduled run) kicks things off. n8n pulls rows from your product sheet so you can control what gets processed by changing a status field or filtering rows.

The product image is fetched from Google Drive and “understood.” The workflow normalizes your Drive URL into a direct link, retrieves the file, then uses OpenAI Vision/Chat to extract useful details (what the product is, what stands out, what should not be changed).

Prompts get generated and organized before any rendering happens. An agent step assembles UGC-style scene prompts using your brand notes, constraints, target model style, aspect ratio, and number of variations. Those prompts are appended into the ad_image sheet, which keeps the review loop clean.

Fal.ai generates the images, then Drive becomes the source of truth. For each prompt, the workflow sends a render request, polls the status until the job completes, downloads the output, uploads it to Google Drive, and updates the corresponding sheet row with the final URL and status.

You can easily modify aspect ratio and prompt style based on your needs. See the full implementation guide below for customization options.

Step-by-Step Implementation Guide

Step 1: Configure the Manual Trigger

Start the workflow with a manual trigger so you can test prompt generation end-to-end before automating it.

Add and keep Manual Start Trigger as the first node.
Leave all fields in Manual Start Trigger at their defaults (no parameters required).
(Optional) Keep Flowpast Branding as a reference sticky note for documentation.

Step 2: Connect Google Sheets Inputs

Pull product rows that need image prompt creation from Google Sheets.

Open Retrieve Sheet Rows and set Document to [YOUR_ID] and Sheet to [YOUR_ID].
Set the filter in Retrieve Sheet Rows to match status equals Create.
Credential Required: Connect your googleSheetsOAuth2Api credentials in Retrieve Sheet Rows.

Tip: Ensure the input sheet columns match the node’s expected fields like product_image_url, product_name, product_description, num_variations, and model_target.

Step 3: Set Up Image Retrieval and AI Analysis

Download the product image and analyze it to generate structured ad insights and prompt plans.

In Fetch Drive File, set Operation to download and File ID to ={{ $json.product_image_url }}.
Credential Required: Connect your googleDriveOAuth2Api credentials in Fetch Drive File.
In Inspect Image Content, keep Resource as image, Input Type as base64, and Operation as analyze.
Set the Text prompt in Inspect Image Content to the provided JSON-guided analysis prompt (as-is).
Credential Required: Connect your openAiApi credentials in Inspect Image Content.
Open Ad Prompt Orchestrator and keep the Text field set to the long JSON instruction that references {{ $('Retrieve Sheet Rows').item.json.* }} values.
Ensure Ad Prompt Orchestrator has Prompt Type set to define and Has Output Parser enabled.
Verify OpenAI Chat Engine is connected as the language model for Ad Prompt Orchestrator; Credential Required: Connect your openAiApi credentials in OpenAI Chat Engine.
Confirm Structured JSON Parser is connected as the output parser for Ad Prompt Orchestrator; add credentials on the parent node (OpenAI Chat Engine), not the parser.

⚠️ Common Pitfall: If Structured JSON Parser is not connected to Ad Prompt Orchestrator, the output won’t be split into scenes for downstream processing.

Step 4: Create Scene Records and Map Product Fields

Split the AI JSON into per-scene items, log them in Sheets, and map image URLs for rendering.

In Explode Scene List, set Field to Split Out to output.scenes.
In Append Sheet Record, set Operation to append and map columns: prompt to ={{ $json.prompt }}, status to Ready, scene_ref to ={{ $('Retrieve Sheet Rows').item.json.product_name }}_{{ $json.scene_id }}, and product_name to ={{ $('Retrieve Sheet Rows').item.json.product_name }}.
Credential Required: Connect your googleSheetsOAuth2Api credentials in Append Sheet Record.
In Map Product Fields, keep Include Other Fields set to true and add the product field expression exactly as provided to convert Drive links into a direct image URL.

Tip: If your Drive URL format is different, update the product expression in Map Product Fields to extract the correct file ID.

Step 5: Render Images via Fal and Poll for Completion

Send prompts to the rendering API, poll for completion, and fetch the final image URL.

In Invoke Fal Image Edit, set URL to =https://queue.fal.run/fal-ai/{{ $('Retrieve Sheet Rows').item.json.model_target }}/edit, Method to =POST, and JSON Body to the provided payload with {{ $json.prompt }} and {{ $json.product }}.
Credential Required: Connect your httpHeaderAuth credentials in Invoke Fal Image Edit.
In Request Image Status, set URL to ={{ $json.status_url }} and keep Authentication as genericCredentialType.
Credential Required: Connect your httpHeaderAuth credentials in Request Image Status.
In Conditional Status Check, set the condition to Left Value ={{ $json.status }} equals Right Value =COMPLETED.
In Delay Cycle, set Amount to 10 to pause before re-checking status.
In Retrieve Render Result, set URL to =https://queue.fal.run/fal-ai/nano-banana/requests/{{ $json.request_id }} and ensure retry is enabled.
Credential Required: Connect your httpHeaderAuth credentials in Retrieve Render Result.
In External Image Fetch, set URL to ={{ $json.images[0].url }} and keep Authentication as genericCredentialType.
Credential Required: Connect your httpHeaderAuth credentials in External Image Fetch.

⚠️ Common Pitfall: If Conditional Status Check never reaches COMPLETED, the loop will keep cycling through Delay Cycle. Verify the Fal API response structure and status values.

Step 6: Save Rendered Images and Update the Sheet

Upload the rendered file to Google Drive and write the output URL back to the tracking sheet.

In Upload to Drive, set Name to ={{ $('Retrieve Sheet Rows').item.json.product_name }}_{{ $('Append Sheet Record').item.json.scene_ref }}.jpeg, Folder to [YOUR_ID], and Input Data Field Name to =data.
Credential Required: Connect your googleDriveOAuth2Api credentials in Upload to Drive.
In Update Sheet Record, set Operation to update and map scene_ref to ={{ $('Append Sheet Record').item.json.scene_ref }}, status to Complete, and output_url to ={{ $json.webViewLink }}.
Credential Required: Connect your googleSheetsOAuth2Api credentials in Update Sheet Record.

Step 7: Test and Activate Your Workflow

Run a manual test, confirm output integrity, then activate for production runs.

Click Execute Workflow on Manual Start Trigger to run the full flow.
Confirm that Append Sheet Record creates rows with Ready status and prompt text.
Verify Upload to Drive creates image files and Update Sheet Record updates output_url and Complete status.
Once verified, toggle the workflow to Active so it can be triggered on demand or swapped to a scheduled trigger later.

🔒

Unlock Full Step-by-Step Guide

Get the complete implementation guide + downloadable template

Common Gotchas

Google Drive permissions matter more than people expect. If Fal.ai can’t fetch your input image, make sure the Drive file is set to “Anyone with link → Viewer” and confirm the workflow is using the converted direct link.
If you’re using Wait nodes or external rendering, processing times vary. Bump up the wait duration if downstream nodes fail on empty responses.
Fal.ai credentials can fail quietly if the header auth is wrong. Check that your Authorization header uses your FAL_KEY environment variable and confirm the status_url is returning a valid job state.

Frequently Asked Questions

How long does it take to set up this ad image automation automation?

About an hour if your Google credentials are ready.

Do I need coding skills to automate ad image automation?

No. You’ll mostly connect accounts and match your Sheet columns to the workflow.

Is n8n free to use for this ad image automation workflow?

Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in OpenAI and Fal.ai API usage costs, which depend on how many images you generate.

Where can I host n8n to run this automation?

Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.

Can I customize this ad image automation workflow for different aspect ratios and styles?

Yes, and it’s one of the best reasons to use this workflow. Update the product row fields like aspect_ratio, constraints, and brand_notes to steer outputs without rewriting prompts by hand. If you want deeper control, adjust the instructions in the Ad Prompt Orchestrator so the generated scene list matches your creative strategy (luxury, cozy, techy, or whatever you’re testing). You can also swap the Fal.ai model in the image generation request if you want faster jobs or a different look.

Why is my Google Drive connection failing in this workflow?

Usually it’s OAuth permissions or the file isn’t actually accessible to the workflow. Reconnect Google Drive in n8n, confirm the account has access to the folder, then verify your input images are shared as “Anyone with link → Viewer” so external services can fetch them. Also double-check that the link is being converted into a direct uc?export=view URL before the image is sent for analysis or generation.

How many images can this ad image automation automation handle?

A lot, as long as you pace it.

Is this ad image automation automation better than using Zapier or Make?

For this workflow, n8n is usually the better fit because you need polling loops (check job status, wait, retry), structured AI output parsing, and more control over how records are written back to Sheets. Those patterns are possible in Zapier or Make, but they get fiddly and can get expensive once you’re looping over lots of variations. n8n also gives you the self-hosting option if you want to run high volume without counting every task. If you only need a simple “new row → create one image → upload” flow, Zapier or Make can be fine. Talk to an automation expert if you want help picking the right approach for your creative volume.

Once this is running, your sheet becomes the command center and Drive becomes the library. The workflow handles the repetitive parts, and you get to focus on what actually moves performance.

Need Help Setting This Up?

Our automation experts can build and customize this workflow for your specific needs. Free 15-minute consultation—no commitment required.

Lisa Granqvist

Workflow Automation Expert

Expert in workflow automation and no-code tools.

{
"id": null,
"meta": {
"instanceId": "f5a8bc42bea862c872bc579aa3ef5c6219f65b1649e4a8c516cdddab3043535b",
"templateCredsSetupCompleted": null,
"templateId": null
},
"name": "Automated Ad Image Prompt Flow",
"tags": [],
"nodes": [
{
"id": "flowpast-topbar-8859",
"name": "Flowpast Branding",
"type": "n8n-nodes-base.stickyNote",
"position": [
0,
20
],
"parameters": {
"color": 7,
"width": 1420,
"height": 80,
"content": "## Flowpast.com | Automation Workflow Library\n**\ud83d\udcd6 Full tutorial & setup guide:** flowpast.com"
},
"typeVersion": 1
},
{
"id": "1c6ff80d-c6dc-485d-9708-38b9f7af4961",
"name": "Manual Start Trigger",
"type": "n8n-nodes-base.manualTrigger",
"position": [
1120,
150
],
"parameters": [],
"typeVersion": 1
},
{
"id": "0ebd375b-d9c4-4ae0-90ac-5295b788a253",
"name": "Retrieve Sheet Rows",
"type": "n8n-nodes-base.googleSheets",
"position": [
930,
180
],
"parameters": {
"options": [],
"filtersUI": {
"values": [
{
"lookupValue": "Create",
"lookupColumn": "status"
}
]
},
"sheetName": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "product"
},
"documentId": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "n8n_Multi_Ad_Image"
}
},
"credentials": {
"googleSheetsOAuth2Api": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 4.7
},
{
"id": "dcb3134c-a3aa-4384-9f37-e59ea93472f7",
"name": "Fetch Drive File",
"type": "n8n-nodes-base.googleDrive",
"position": [
720,
120
],
"parameters": {
"fileId": {
"__rl": true,
"mode": "url",
"value": "={{ $json.product_image_url }}"
},
"options": [],
"operation": "download"
},
"credentials": {
"googleDriveOAuth2Api": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 3
},
{
"id": "e982d7e4-dced-4897-85cf-708e09d5bc4e",
"name": "Inspect Image Content",
"type": "@n8n/n8n-nodes-langchain.openAi",
"position": [
520,
170
],
"parameters": {
"text": "=You are an AI system specialized in **image analysis for advertising**. \nYour task is to carefully analyze the given product image and generate \nstructured insights that will later be used to create high-quality \nadvertising prompts.\n\n### Instructions:\n1. **Identify Product**  \n   - Type of product (e.g., sneakers, skincare, electronics).  \n   - Key visible features (color, shape, size, packaging).  \n   - Materials or textures that can be inferred.  \n\n2. **Branding & Style**  \n   - Any visible logos, brand elements, or design patterns.  \n   - Style impression (luxury, casual, eco-friendly, tech-oriented).  \n\n3. **Target Audience Guess**  \n   - Suggest who the product might appeal to (young adults, parents, professionals, etc.).  \n\n4. **Ad Creative Angles (3\u20135 ideas)**  \n   - Marketing hooks inspired by the image.  \n   - Focus on visual storytelling (e.g., \u201cminimalist lifestyle shot\u201d, \u201cdynamic sports action scene\u201d).  \n\n5. **Output Format (JSON)**  \n   ```json\n   {\n     \"product_type\": \"...\",\n     \"features\": [\"...\", \"...\"],\n     \"style\": \"...\",\n     \"target_audience\": \"...\",\n     \"ad_angles\": [\n       \"Ad idea 1\",\n       \"Ad idea 2\",\n       \"Ad idea 3\",\n       \"Ad idea 4\",\n       \"Ad idea 5\"\n     ]\n   }\n",
"modelId": {
"__rl": true,
"mode": "list",
"value": "gpt-4o-mini",
"cachedResultName": "GPT-4O-MINI"
},
"options": [],
"resource": "image",
"inputType": "base64",
"operation": "analyze"
},
"credentials": {
"openAiApi": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 1.8
},
{
"id": "c87947e9-97eb-4758-b18c-bf14cc232573",
"name": "Ad Prompt Orchestrator",
"type": "@n8n/n8n-nodes-langchain.agent",
"position": [
320,
140
],
"parameters": {
"text": "=You will receive a real product image and its short description. \nAnalyze them and generate exactly {{ $('Retrieve Sheet Rows').item.json.num_variations }} distinct, production-ready advertising image prompts.\nIf a campaign is provided, weave its seasonal or promotional context into the prompts naturally. \nIf campaign is empty or \"none\", ignore it.\n\n=== INPUTS ===\n- Product Image URL: {{ $('Retrieve Sheet Rows').item.json.product_image_url }}\n- Product Name: {{ $('Retrieve Sheet Rows').item.json.product_name }}\n- Product Description: \"{{ $('Retrieve Sheet Rows').item.json.product_description }}\"\n- Campaign (optional): {{ $('Retrieve Sheet Rows').item.json.campaign }}                      // e.g., \"8.8\", \"10.10\", \"Songkran\", \"Halloween\", \"none\"\n- Target Model: {{ $('Retrieve Sheet Rows').item.json.model_target }}                    // e.g., \"SDXL\", \"Midjourney\", \"Flux\", \"Veo\"\n- Aspect Ratio:  {{ $('Retrieve Sheet Rows').item.json.aspect_ratio }}                     // e.g., \"1:1\", \"4:5\", \"16:9\"\n- Number of Variations (N): {{ $('Retrieve Sheet Rows').item.json.num_variations }}\n- Brand Guidelines (optional): {{ $('Retrieve Sheet Rows').item.json.brand_notes }}           // e.g., brand palette hex codes, keep logo visible, avoid surreal\n- Creative Constraints (optional): {{ $('Retrieve Sheet Rows').item.json.constraints }}       // e.g., \"no people\", \"no surreal\", \"text-safe negative space\"\n\n=== TASK ===\n1) Treat the product image as ground truth; keep product identity accurate (color, form, logo).\n2) Generate {{ $('Retrieve Sheet Rows').item.json.num_variations }} visually distinct prompts that are advertising-ready and usable by {{ $('Retrieve Sheet Rows').item.json.model_target }}.\n3) Vary at least one of: scene/environment, style, lighting, camera, background, composition, color mood, post-process.\n4) If \"campaign\" is provided (not empty/none), incorporate tasteful thematic elements: \n   - Halloween \u2192 spooky mood, pumpkins, warm/orange accents, moody lighting\n   - Songkran \u2192 water splashes, summer vibe, bright daylight, outdoor festivity\n   - 8.8 / 10.10 / 11.11 \u2192 energetic sale vibe, dynamic composition, bold accents\n   - Christmas \u2192 warm lights, pine elements, cozy red/green palette\n   Keep it subtle and brand-safe; do not add new logos or misleading claims.\n5) Follow strictly the JSON schema in the **system prompt** (no extra text).\n6) If any subfield cannot be inferred safely, leave it as an empty string \"\".\n7) Ensure each \"prompt\" is **one line** and does not break JSON.\n\n=== REMINDERS ===\n- Avoid vague adjectives. Be concrete and visual.\n- Composition tips: centered hero, rule-of-thirds lifestyle, top-down flat-lay, negative space for copy.\n- Lighting tips: softbox/rim for packshots, natural daylight for lifestyle, golden hour for warmth.\n- Camera tips: 35mm/50mm for natural look, 85mm for product portrait with shallow DoF, macro for texture.\n- Color mood: align gently with any campaign; otherwise, use clean brand-safe palettes.\n\n=== RETURN FORMAT ===\n- Return a **single JSON object** only, matching the schema in the system prompt.\n",
"options": {
"systemMessage": "=You are an **Advertising Image Prompt Designer**.\n\nInputs you may receive:\n- product_image_url (image of the real product; treat as ground truth)\n- product_name\n- product_description\n- campaign (optional; e.g., \"8.8\", \"10.10\", \"Songkran\", \"Halloween\", \"Black Friday\", \"Christmas\")\n- model_target (e.g., \"SDXL\", \"Midjourney\", \"Flux\", \"Veo\")\n- aspect_ratio (e.g., \"1:1\", \"4:5\", \"16:9\")\n- N (number of distinct variations to generate)\n\nGoal:\n- Analyze the given product image, short description, and optional campaign context.\n- Produce **N distinct, production-ready image prompts** for ad creation.\n- **Preserve the product\u2019s true identity** (color, form, logo, key features). Do not invent or alter the product.\n\nOutput:\n- Return **JSON only** following the schema below. **No extra text** outside the JSON.\n\nQuality rules:\n- Each prompt must be visually distinct (vary at least one major axis: scene, mood, style, lighting, camera, background, composition).\n- Use concise, concrete descriptors; avoid vague words (e.g., \u201cnice\u201d, \u201cbeautiful\u201d).\n- Keep prompts brand-safe and realistic unless the user explicitly allows surreal concepts.\n- Prefer modern ad aesthetics: clean hero shots, lifestyle, editorial, macro texture, flat-lay, outdoor ambient, etc.\n- If any requested field is missing in the user input, **infer from the image**; if you cannot infer safely, set that subfield to an empty string.\n- If a **campaign** is provided, incorporate its seasonal or promotional theme **naturally and tastefully** (e.g.,  \n  \u2022 Halloween \u2192 moody/spooky, pumpkins, warm/orange accents, fog  \n  \u2022 Songkran \u2192 water splash, festive Thai summer vibe, bright daylight  \n  \u2022 8.8 / 10.10 / 11.11 \u2192 bold sale energy, dynamic composition, price-tag visual cues without adding logos  \n  \u2022 Christmas \u2192 cozy, pine, warm lights, red/green palette  \n  \u2022 Black Friday \u2192 high contrast, dark sleek backdrops).  \n  If no campaign is provided (empty or \u201cnone\u201d), **ignore** campaign styling.\n\nTarget models:\n- Prompts should be usable by **Midjourney, SDXL, Flux, or Veo**. Avoid model-specific tokens unless provided in the user input (e.g., \u201c--ar 16:9\u201d, \u201c--style raw\u201d).\n\nPrompt composition guidance (recommended ingredients):\n- **Subject/Product**: exact product + salient features visible in the image (material, finish, color, unique parts).\n- **Scene/Environment**: studio / lifestyle / outdoor / occasion-linked (if campaign present).\n- **Style/Art direction**: minimalist / cinematic / editorial / lifestyle / high-fashion / tech-forward / cozy, etc.\n- **Lighting**: softbox / rim light / natural daylight / golden hour / neon / HDRI; mention shadows/reflections if relevant.\n- **Camera**: 35mm / 85mm / wide / macro / top-down / tilt-shift, F-stop or DoF if helpful.\n- **Composition**: centered / rule of thirds / flat-lay / hero shot / negative space for text.\n- **Background**: seamless white/black, textured concrete, gradient, seasonal set pieces (only if campaign present).\n- **Color mood**: pastel / vibrant / monochrome / brand palette; align gently with any campaign palette.\n- **Post-process**: subtle film grain / glossy reflections / HDR / product retouch / shadow catching.\n- **Marketing angle**: premium luxury / eco-friendly / active lifestyle / tech-forward / festive sale energy.\n\nBrand-safety & integrity:\n- Do **not** add or modify logos, certifications, or claims not present in the description.\n- Do **not** imply medical/health claims unless explicitly provided.\n- Avoid depicting real people/celebrities. If humans are implied, keep them generic and non-identifiable.\n\nRequired JSON schema (return this object only):\n{\n  \"product_name\": string,\n  \"product_description\": string,\n  \"image_url\": string,\n  \"campaign\": string,                   // optional; leave \"\" if none\n  \"model_target\": string,               // e.g., \"SDXL\", \"Midjourney\", \"Flux\", \"Veo\"\n  \"aspect_ratio\": string,               // e.g., \"1:1\", \"4:5\", \"16:9\"\n  \"variations\": [\n    {\n      \"variation_id\": number,\n      \"prompt\": string,                 // full generation prompt (one line)\n      \"style\": string,                  // e.g., minimalist / cinematic / editorial / lifestyle / surreal\n      \"scene\": string,                  // e.g., studio / outdoor / cafe / gym / beach\n      \"lighting\": string,               // e.g., softbox / natural daylight / neon / golden hour / rim light\n      \"camera\": string,                 // e.g., 35mm / 85mm / macro / top-down / tilt-shift, include DoF if useful\n      \"composition\": string,            // e.g., centered / rule of thirds / flat-lay / hero shot / negative space\n      \"background\": string,             // e.g., white seamless / textured concrete / gradient / city night\n      \"color_mood\": string,             // e.g., pastel / vibrant / monochrome / brand palette\n      \"post_process\": string,           // e.g., subtle film grain / glossy reflections / HDR / product retouch\n      \"marketing_angle\": string         // e.g., premium luxury / eco-friendly / active lifestyle / tech-forward\n    }\n  ]\n}\n\nValidation:\n- Produce **exactly N** variations.\n- Ensure each \"prompt\" is a **single line** without JSON-breaking characters.\n- Escape any internal quotes if needed. Do not include commentary outside the JSON block.\n\nFailure handling:\n- If the image cannot be analyzed (unreachable URL or ambiguous), still return a valid JSON with minimal safe prompts derived from the description; leave unknown subfields as \"\" and add a short note in \"notes\".\n"
},
"promptType": "define",
"hasOutputParser": true
},
"typeVersion": 2.2
},
{
"id": "f4b60076-6dd2-4da1-90db-117eabe4715c",
"name": "OpenAI Chat Engine",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
360,
280
],
"parameters": {
"model": {
"__rl": true,
"mode": "list",
"value": "gpt-4.1-mini"
},
"options": []
},
"credentials": {
"openAiApi": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 1.2
},
{
"id": "bc750bdb-45ad-4380-ad1d-6688cb242850",
"name": "Structured JSON Parser",
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"position": [
140,
300
],
"parameters": {
"jsonSchemaExample": "\n{\n  \"product_name\": \"Smart Water Bottle\",\n  \"product_description\": \"A sleek stainless steel smart water bottle with LED temperature display.\",\n  \"image_url\": \"https://drive.google.com/uc?export=view&id=[YOUR_ID]\",\n  \"scenes\": [\n    {\n      \"scene_id\": 1,\n      \"prompt\": \"smart stainless steel water bottle on a minimalist white marble counter, cinematic studio lighting, soft shadows, product hero shot, centered composition\"\n    },\n    {\n      \"scene_id\": 2,\n      \"prompt\": \"smart water bottle on a wooden gym bench, fitness lifestyle, natural sunlight from window, sweat towel nearby, cinematic depth of field\"\n    },\n    {\n      \"scene_id\": 3,\n      \"prompt\": \"product placed in a futuristic cyberpunk neon city background, glowing reflections, dramatic rim lighting, 35mm lens, bold composition\"\n    }\n  ]\n}\n"
},
"typeVersion": 1.3
},
{
"id": "12f40103-c60f-4a7b-82d1-960bd6b32755",
"name": "Explode Scene List",
"type": "n8n-nodes-base.splitOut",
"position": [
0,
180
],
"parameters": {
"options": [],
"fieldToSplitOut": "output.scenes"
},
"typeVersion": 1
},
{
"id": "e695e646-9759-46c4-a33c-36cda24da934",
"name": "Append Sheet Record",
"type": "n8n-nodes-base.googleSheets",
"position": [
220,
360
],
"parameters": {
"columns": {
"value": {
"prompt": "={{ $json.prompt }}",
"status": "Ready",
"scene_ref": "={{ $('Retrieve Sheet Rows').item.json.product_name }}_{{ $json.scene_id }}",
"product_name": "={{ $('Retrieve Sheet Rows').item.json.product_name }}"
},
"schema": [
{
"id": "scene_ref",
"type": "string",
"display": true,
"required": false,
"displayName": "scene_ref",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "product_name",
"type": "string",
"display": true,
"required": false,
"displayName": "product_name",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "prompt",
"type": "string",
"display": true,
"required": false,
"displayName": "prompt",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "status",
"type": "string",
"display": true,
"required": false,
"displayName": "status",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "output_url",
"type": "string",
"display": true,
"removed": true,
"required": false,
"displayName": "output_url",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [],
"attemptToConvertTypes": false,
"convertFieldsToString": false
},
"options": [],
"operation": "append",
"sheetName": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "ad_image"
},
"documentId": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "n8n_Multi_Ad_Image"
}
},
"credentials": {
"googleSheetsOAuth2Api": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 4.7
},
{
"id": "69afdf78-491a-4952-936e-03208adc6787",
"name": "Map Product Fields",
"type": "n8n-nodes-base.set",
"position": [
440,
340
],
"parameters": {
"options": [],
"assignments": {
"assignments": [
{
"id": "62a409c6-c607-452b-9bbb-f3d29c86dddf",
"name": "product",
"type": "string",
"value": "={{ \n  (() => {     \n    const u = $('Retrieve Sheet Rows').item.json.product_image_url || '';     \n    const q = u.match(/[?&]id=([-\\w]{25,})/);     \n    const d = u.match(/\\/d\\/([-\\w]{25,})/);     \n    const any = u.match(/[-\\w]{25,}/);     \n    const id = q?.[1] || d?.[1] || (any ? any[0] : '');     \n    return id ? 'https://drive.google.com/uc?export=view&id=' + id : '';   \n  })() \n}}"
}
]
},
"includeOtherFields": true
},
"typeVersion": 3.4
},
{
"id": "16e46dfb-8cf7-45fd-9d72-c71b1b3eefad",
"name": "Invoke Fal Image Edit",
"type": "n8n-nodes-base.httpRequest",
"position": [
660,
380
],
"parameters": {
"url": "=https://queue.fal.run/fal-ai/{{ $('Retrieve Sheet Rows').item.json.model_target }}/edit",
"method": "=POST",
"options": [],
"jsonBody": "={\n     \"prompt\": \"{{ $json.prompt }} {{ $json.notes }}\",\n     \"image_urls\": [\n       \"{{ $json.product }}\"\n     ],\n     \"num_images\": 1,\n     \"output_format\": \"jpeg\"\n}",
"sendBody": true,
"specifyBody": "json",
"authentication": "genericCredentialType",
"genericAuthType": "httpHeaderAuth"
},
"credentials": {
"httpHeaderAuth": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 4.2
},
{
"id": "b313b28e-10e0-4815-b8e5-a5e8c1ad4542",
"name": "Request Image Status",
"type": "n8n-nodes-base.httpRequest",
"position": [
880,
330
],
"parameters": {
"url": "={{ $json.status_url }}",
"options": [],
"authentication": "genericCredentialType",
"genericAuthType": "httpHeaderAuth"
},
"credentials": {
"httpHeaderAuth": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 4.2
},
{
"id": "c75b4d0b-3309-4f03-8402-d13accc7edbc",
"name": "Conditional Status Check",
"type": "n8n-nodes-base.if",
"position": [
1100,
380
],
"parameters": {
"options": [],
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "38f5427f-4fc7-4c07-87d5-fa7f5964deb2",
"operator": {
"name": "filter.operator.equals",
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $json.status }}",
"rightValue": "=COMPLETED"
}
]
}
},
"typeVersion": 2.2
},
{
"id": "3c27ccb4-b7f0-4ddd-ad17-bb3563920552",
"name": "Retrieve Render Result",
"type": "n8n-nodes-base.httpRequest",
"position": [
1320,
330
],
"parameters": {
"url": "=https://queue.fal.run/fal-ai/nano-banana/requests/{{ $json.request_id }}",
"options": [],
"authentication": "genericCredentialType",
"genericAuthType": "httpHeaderAuth"
},
"credentials": {
"httpHeaderAuth": {
"id": "credential-id",
"name": ""
}
},
"retryOnFail": true,
"typeVersion": 4.2,
"alwaysOutputData": false
},
{
"id": "ee86e519-9ea5-41f0-b887-559558b1b206",
"name": "External Image Fetch",
"type": "n8n-nodes-base.httpRequest",
"position": [
1080,
560
],
"parameters": {
"url": "={{ $json.images[0].url }}",
"options": [],
"authentication": "genericCredentialType",
"genericAuthType": "httpHeaderAuth"
},
"credentials": {
"httpHeaderAuth": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 4.2
},
{
"id": "8ae70d5b-3dd5-42e6-9c32-f6d36d299b21",
"name": "Upload to Drive",
"type": "n8n-nodes-base.googleDrive",
"position": [
840,
600
],
"parameters": {
"name": "={{ $('Retrieve Sheet Rows').item.json.product_name }}_{{ $('Append Sheet Record').item.json.scene_ref }}.jpeg",
"driveId": {
"__rl": true,
"mode": "list",
"value": "My Drive"
},
"options": [],
"folderId": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "outputImage"
},
"inputDataFieldName": "=data"
},
"credentials": {
"googleDriveOAuth2Api": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 3
},
{
"id": "56a8079f-6e4d-46db-b975-433463b9fa68",
"name": "Update Sheet Record",
"type": "n8n-nodes-base.googleSheets",
"position": [
600,
560
],
"parameters": {
"columns": {
"value": {
"status": "Complete",
"scene_ref": "={{ $('Append Sheet Record').item.json.scene_ref }}",
"output_url": "={{ $json.webViewLink }}"
},
"schema": [
{
"id": "scene_ref",
"type": "string",
"display": true,
"removed": false,
"required": false,
"displayName": "scene_ref",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "product_name",
"type": "string",
"display": true,
"removed": true,
"required": false,
"displayName": "product_name",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "prompt",
"type": "string",
"display": true,
"removed": true,
"required": false,
"displayName": "prompt",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "notes",
"type": "string",
"display": true,
"removed": true,
"required": false,
"displayName": "notes",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "status",
"type": "string",
"display": true,
"removed": false,
"required": false,
"displayName": "status",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "output_url",
"type": "string",
"display": true,
"required": false,
"displayName": "output_url",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "row_number",
"type": "number",
"display": true,
"removed": true,
"readOnly": true,
"required": false,
"displayName": "row_number",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [
"scene_ref"
],
"attemptToConvertTypes": false,
"convertFieldsToString": false
},
"options": [],
"operation": "update",
"sheetName": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "ad_image"
},
"documentId": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "n8n_Multi_Ad_Image"
}
},
"credentials": {
"googleSheetsOAuth2Api": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 4.7
},
{
"id": "d53e1592-7956-4154-9df9-bf8b7943e7fc",
"name": "Delay Cycle",
"type": "n8n-nodes-base.wait",
"position": [
740,
620
],
"webhookId": "98549686-33c2-4bef-a176-92ead56ad0d1",
"parameters": {
"amount": 10
},
"typeVersion": 1.1
}
],
"active": false,
"pinData": [],
"settings": {
"executionOrder": "v1"
},
"versionId": null,
"connections": {
"Conditional Status Check": {
"main": [
[
{
"node": "Retrieve Render Result",
"type": "main",
"index": 0
}
],
[
{
"node": "Delay Cycle",
"type": "main",
"index": 0
}
]
]
},
"Delay Cycle": {
"main": [
[
{
"node": "Request Image Status",
"type": "main",
"index": 0
}
]
]
},
"Ad Prompt Orchestrator": {
"main": [
[
{
"node": "Explode Scene List",
"type": "main",
"index": 0
}
]
]
},
"Explode Scene List": {
"main": [
[
{
"node": "Append Sheet Record",
"type": "main",
"index": 0
}
]
]
},
"Map Product Fields": {
"main": [
[
{
"node": "Invoke Fal Image Edit",
"type": "main",
"index": 0
}
]
]
},
"Upload to Drive": {
"main": [
[
{
"node": "Update Sheet Record",
"type": "main",
"index": 0
}
]
]
},
"External Image Fetch": {
"main": [
[
{
"node": "Upload to Drive",
"type": "main",
"index": 0
}
]
]
},
"Inspect Image Content": {
"main": [
[
{
"node": "Ad Prompt Orchestrator",
"type": "main",
"index": 0
}
]
]
},
"Fetch Drive File": {
"main": [
[
{
"node": "Inspect Image Content",
"type": "main",
"index": 0
}
]
]
},
"Retrieve Render Result": {
"main": [
[
{
"node": "External Image Fetch",
"type": "main",
"index": 0
}
]
]
},
"Request Image Status": {
"main": [
[
{
"node": "Conditional Status Check",
"type": "main",
"index": 0
}
]
]
},
"OpenAI Chat Engine": {
"ai_languageModel": [
[
{
"node": "Ad Prompt Orchestrator",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Append Sheet Record": {
"main": [
[
{
"node": "Map Product Fields",
"type": "main",
"index": 0
}
]
]
},
"Retrieve Sheet Rows": {
"main": [
[
{
"node": "Fetch Drive File",
"type": "main",
"index": 0
}
]
]
},
"Structured JSON Parser": {
"ai_outputParser": [
[
{
"node": "Ad Prompt Orchestrator",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"Invoke Fal Image Edit": {
"main": [
[
{
"node": "Request Image Status",
"type": "main",
"index": 0
}
]
]
},
"Manual Start Trigger": {
"main": [
[
{
"node": "Retrieve Sheet Rows",
"type": "main",
"index": 0
}
]
]
}
}
}