Auto-Generate Product Photos and Videos from a Reference Image with n8n

User submits a product photo and description → GPT-4.1 rewrites the prompt → OpenAI generates a new image → RunwayML renders a video → Gmail sends the results automatically

Who

E-commerce operators, content creators, marketing teams

You get

Submit a reference product image and automatically receive an AI-generated product photo and a 5-second marketing video

You need

n8n · google-drive · gmail · runwayml · imgbb

Steps

1. Configure the form trigger

   Add the "A.I Stunning Generator Interface" Form Trigger node in n8n. Set the form title to "Stelo Link Studio" and add four required fields: Image File (file, accepts .jpeg/.jpg/.png/.webm), Product Image Name (text, placeholder "Name of New Image"), Prompt A.I (textarea, placeholder "Explain your Vision. Think [Background, Content and Styles]"), and Email (email, placeholder "[email protected]"). Once the workflow is active, n8n generates a public form URL that you can share with submitters.

2. Back up the original image to Google Drive

   After form submission, upload the original image to Google Drive for backup. Add an "Upload to Google Drive" node using Google Drive OAuth2 credentials. Set the filename to "={{ $json['Product Image Name'] }} (original)", replace folderId with your own Google Drive folder ID (copied from the folder URL), and set inputDataFieldName to "Image_File". Then add a "Download from Google Drive" node to download the file by the returned id for use in subsequent steps.

3. Rewrite prompt with AI and generate a new image

   Add an "AI Agent" node connected to the "Chat GPT 4.1" language model (select gpt-4.1). Set the AI Agent input text to: "=Product : {{ $('A.I Stunning Generator Interface').item.json['Product Image Name'] }}\nDescription : {{ $('A.I Stunning Generator Interface').item.json['Prompt A.I'] }}". Paste the following system prompt (verbatim from workflow.json) into the AI Agent's System Message field so GPT-4.1 rewrites the user's description into a professional image generation prompt. Then add an "OpenAi Create Image" node calling https://api.openai.com/v1/images/edits with HTTP Header Auth (Bearer YOUR_OPENAI_API_KEY), multipart/form-data body: image field from the previous download's data, prompt field from AI Agent output "={{ $('AI Agent').item.json.output }}", model set to gpt-image-1. Finally, use a "Convert to File" node to convert the returned data[0].b64_json into a binary file.

   Execution prompt · Source text。AI Agent node System Message field

   Copy prompt

   =# Overview
   You are a world-class marketing strategist and an expert text-to-image prompt engineer
   specializing in creating hyper-realistic, high-quality product photography prompts for AI image
   generation models.
   ## Your Objective:
   When given a product description, your task is to craft a detailed, professional prompt that
   results in a hyper-realistic, clean, and visually stunning product image suitable for marketing
   material, advertising campaigns, or e-commerce platforms.
   ## Key Requirements:
   - Focus on hyper-realistic, professional studio photography.
   - Maintain a clean, minimalistic, and elegant visual style.
   - Highlight the product as the main subject with sharp detail and perfect lighting.
   - Use backgrounds that complement but do not overpower the product (e.g., soft gradient,
   light-colored, or pure white).
   - Include professional lighting details such as "softbox lighting"
   ,
   "studio lights"
   , or "natural soft
   shadows"
   .
   - Emphasize realism, texture, and color accuracy.
   - Maintain a high-end, premium look and feel.
   - If applicable, suggest a subtle setting that enhances the product's story (e.g., a luxury kitchen
   counter for a high-end blender).
   ## Output Format:
   - Write a single text-to-image prompt optimized for a professional AI image model.
   - Be direct and descriptive without using excessive words.
   - Avoid unnecessary repetition or adjectives that do not enhance the image quality.
   - Ensure the prompt is complete and ready for direct input into an AI model.
   ## Tone:
   Professional, precise, clean, and optimized for maximum realism and marketing impact.
   # Example Input:
   "A premium wireless Bluetooth speaker in matte black, cylindrical shape, modern design.
   "
   # Example Output:
   "Hyper-realistic product photo of a premium matte black wireless Bluetooth speaker with a
   cylindrical modern design, centered on a clean white studio background, softbox lighting with
   natural shadows, sharp focus on texture and material, minimalistic, professional advertising
   shot.
   "

4. Upload image to get a public URL

   Add a "Get URL" node calling the ImgBB upload API (https://api.imgbb.com/1/upload) with HTTP Query Auth (parameter name key, value YOUR_IMGBB_API_KEY). Request type multipart/form-data; body parameter image set to formBinaryData with inputDataFieldName "data" (from the previous Convert to File output). After successful upload, ImgBB returns data.url for the next step.

5. Submit video generation and poll for completion

   Add a "Create Video" node sending a POST request to https://api.dev.runwayml.com/v1/image_to_video with HTTP Header Auth (Bearer YOUR_RUNWAYML_API_KEY). The header must include X-Runway-Version: 2024-11-06 (do not change). Body parameters: promptImage from previous ImgBB data.url, model gen4_turbo, promptText fixed video description (see prompt below), duration 5, ratio 720:1280. Then add a "60 Seconds" Wait node (60s), followed by an "HTTP Request" polling node with URL "=https://api.dev.runwayml.com/v1/tasks/{{ $('Create Video').item.json.id }}" and the same X-Runway-Version: 2024-11-06 header. Add an "If" node checking whether the returned status equals "SUCCEEDED" or "RUNNING": if SUCCEEDED, proceed to email; otherwise enter a "Wait" node for 30 seconds and loop back to the HTTP Request.

   Execution prompt · Source text。Create Video node bodyParameters.promptText field

   Copy prompt

   =Create a highly professional marketing video from the provided product photo. Simulate a smooth 360 rotation, realistic 3D turntable rotation around the product, as if the product is slowly rotating in place. The movement should be continuous, slow, and elegant — no sudden pans, jerks, or camera cuts. Always keep the entire product fully in frame, centered, and clearly visible at all times. Avoid zooming in or cropping. Focus on a premium, clean, and modern aesthetic that feels suitable for commercial marketing materials. No flashy effects, transitions, or overlays — only a subtle, realistic 3D rotation that highlights the product in the most polished way possible.

6. Send the results via email

   Add a "Send a message" Gmail node using Gmail OAuth2 credentials (the sending account is the authorized Google account). Set sendTo to "={{ $('A.I Stunning Generator Interface').item.json.Email }}", subject to "=Marketing Material : {{ $('A.I Stunning Generator Interface').item.json['Product Image Name'] }}", emailType text, and message body as follows (verbatim from workflow.json). Disable appendAttribution.

   Execution prompt · Source text。Send a message node message field

   Copy prompt

   =Hey!
   
   Here is your photo: {{ $('Get URL').item.json.data.url_viewer }}
   
   and your video: {{ $json.output[0] }}
   Cheers
   Thank you for choosing Stelo Link

7. Configure credentials and permissions

   Configure the following five credentials in n8n's Credentials page per the source instructions: 1) OpenAI — HTTP Header Auth, Header name: Authorization, Value: Bearer YOUR_OPENAI_API_KEY, connect to Chat GPT 4.1 and OpenAi Create Image nodes; 2) Google Drive — Google Drive OAuth2, authorize via OAuth flow, then update the Folder ID in the Upload to Google Drive node; 3) ImgBB — HTTP Query Auth, Parameter name: key, Value: YOUR_IMGBB_API_KEY, free signup at imgbb.com, connect to Get URL node; 4) RunwayML — HTTP Header Auth, Header name: Authorization, Value: Bearer YOUR_RUNWAYML_API_KEY, connect to Create Video and HTTP Request polling nodes; 5) Gmail — Gmail OAuth2, authorize via OAuth flow, connect to Send a Message node.

You are setting up an n8n workflow: a user submits a product reference image and description, and the system automatically generates an AI product photo and marketing video, then emails the results. Please confirm you have ready: an n8n instance, an OpenAI API Key with gpt-image-1 access, a Google Drive folder, an ImgBB key, a RunwayML key, and a Gmail OAuth account. Then import workflow.json or rebuild the nodes step by step.