Creating branded AI models at scale has become essential for modern content pipelines. With the rise of model-centric marketing, brands increasingly need consistent output, reusable prompts, and automated storage. Fortunately, n8n makes it possible to build this entire pipeline without writing a single line of backend code.
In this guide, we will walk through a complete n8n AI Model Workflow that automatically:
- Generates an AI model image from a custom prompt
- Enhances the image and converts it to PNG
- Uploads it to Catbox (free image hosting)
- Saves the image URL and metadata to Airtable
Throughout this tutorial, we will also cover how to customize your prompts for different model styles and how the workflow functions from end to end.
Why Use This n8n AI Model Workflow?
This workflow is designed for creators, brands, and AI enthusiasts who want:
- Consistent AI model generation
- A library of reusable brand models
- Instant hosting via Catbox
- Automatic organization using Airtable
- A no-code, repeatable process
As a result, you can create multiple models in minutes instead of building images manually.
Workflow Overview
Before we examine the nodes individually, here is what the pipeline accomplishes:
- Manual Trigger starts the workflow.
- Model Specifications sets the model’s name and prompt.
- Generate Model Image sends the prompt to FLUX.1-dev on Hugging Face.
- Edit Image ensures proper size, resolution, and PNG formatting.
- Upload to Catbox hosts the image online.
- Save to Airtable stores model name, image URL, and timestamp.
This structure makes the process scalable and easy to modify.
Generic Prompt Example (You Can Swap This Easily)
To generate different AI models, you can use a baseline prompt like:
“High-quality AI fashion model, studio lighting, clean background, full-body shot, natural pose, realistic skin texture, photorealistic, professional editorial photography.”
From there, you can adjust the style:
- Traditional
- Professional
- Modern
- Luxury
- Streetwear
- Fitness
This flexibility makes the workflow ideal for brand-specific modeling systems.
Step-by-Step Breakdown of the Workflow
1. Manual Trigger
This node simply gives you an easy button inside n8n to start the workflow. Nothing needs to be configured here.
2. Model Specifications (Set Node)
This node defines two key variables:
- model_name
- prompt
By editing this single node, you can generate many different models. Additionally, n8n passes these values dynamically to all future nodes.
3. Generate Model Image (HTTP Request)
This step sends your prompt to FLUX.1-dev, one of the strongest open-source diffusion models available.
- Method: POST
- URL:
https://router.huggingface.co/hf-inference/models/black-forest-labs/FLUX.1-dev - Headers: JSON request, PNG response
- Body: Prompt + parameters
- Output: Raw AI-generated image
The workflow requests an image with:
- 28 inference steps
- A guidance scale of 3.5
These settings offer a balance between speed and quality.
4. Edit Image
The “Edit Image” node:
- Converts the output to PNG
- Resizes to 1024×1024 for consistency
- Names the file based on the model
This produces a clean, standardized asset ready for hosting.
5. Upload to Catbox
Catbox.moe is a fast, reliable, and free image-hosting service.
The workflow:
- Uploads the PNG
- Retrieves the hosted URL
- Passes it to Airtable
Because Catbox requires no API key, this step is easy to configure.
6. Save to Airtable
The final node stores:
- Model Name
- Image URL
- Timestamp
This forms a growing database of AI-generated models you can reference anytime.
Full Workflow JSON (Cleaned and Safe)
Below is the complete workflow with all sensitive credentials removed. You can import it directly into n8n.
{
"nodes": [
{
"parameters": {},
"id": "manual-1",
"name": "Manual Trigger",
"type": "n8n-nodes-base.manualTrigger",
"position": [240, 304],
"typeVersion": 1
},
{
"parameters": {
"assignments": {
"assignments": [
{
"name": "model_name",
"type": "string",
"value": "Sample Model"
},
{
"name": "prompt",
"type": "string",
"value": "High-quality AI fashion model..."
}
]
}
},
"id": "set-model-1",
"name": "Model Specifications",
"type": "n8n-nodes-base.set",
"position": [448, 304],
"typeVersion": 3.4
},
{
"parameters": {
"method": "POST",
"url": "https://router.huggingface.co/hf-inference/models/black-forest-labs/FLUX.1-dev",
"sendHeaders": true,
"headerParameters": {
"parameters": [
{"name": "Content-Type", "value": "application/json"},
{"name": "Accept", "value": "image/png"}
]
},
"sendBody": true,
"specifyBody": "json",
"jsonBody": "={ \"inputs\": \"{{ $json.prompt }}\", \"parameters\": { \"num_inference_steps\": 28, \"guidance_scale\": 3.5 } }",
"options": {
"response": { "response": { "responseFormat": "file" } },
"timeout": 180000
}
},
"id": "http-flux-1",
"name": "Generate Model Image",
"type": "n8n-nodes-base.httpRequest",
"position": [640, 304],
"typeVersion": 4.2
},
{
"parameters": {
"operation": "resize",
"width": 1024,
"height": 1024,
"options": {
"fileName": "={{ $json.model_name.toLowerCase().replace(/[^a-z0-9]/g, '-') }}.png",
"format": "png"
}
},
"type": "n8n-nodes-base.editImage",
"position": [848, 304],
"id": "edit-image-1",
"name": "Edit Image"
},
{
"parameters": {
"method": "POST",
"url": "https://catbox.moe/user/api.php",
"sendBody": true,
"contentType": "multipart-form-data",
"bodyParameters": {
"parameters": [
{ "name": "reqtype", "value": "fileupload" },
{ "parameterType": "formBinaryData", "name": "fileToUpload", "inputDataFieldName": "data" }
]
}
},
"id": "upload-catbox-1",
"name": "Upload to Catbox",
"type": "n8n-nodes-base.httpRequest",
"position": [1024, 304],
"typeVersion": 4.2
},
{
"parameters": {
"operation": "create",
"base": { "mode": "id", "value": "YOUR_AIRTABLE_BASE_ID" },
"table": { "mode": "id", "value": "YOUR_AIRTABLE_TABLE_ID" },
"columns": {
"value": {
"Model Name": "={{ $('Model Specifications').item.json.model_name }}",
"Model Image": "={{ [{ url: $('Upload to Catbox').item.json.data }] }}",
"Created Date": "={{ $now.toISO() }}"
}
}
},
"id": "airtable-1",
"name": "Save to Airtable",
"type": "n8n-nodes-base.airtable",
"position": [1216, 304],
"typeVersion": 2.1
}
],
"connections": {
"Manual Trigger": { "main": [[{ "node": "Model Specifications" }]] },
"Model Specifications": { "main": [[{ "node": "Generate Model Image" }]] },
"Generate Model Image": { "main": [[{ "node": "Edit Image" }]] },
"Edit Image": { "main": [[{ "node": "Upload to Catbox" }]] },
"Upload to Catbox": { "main": [[{ "node": "Save to Airtable" }]] }
}
}
Best Practices for Improving Model Quality
To elevate results, consider adding:
- “visible fingers”
- “studio lighting”
- “high-resolution skin texture”
- “professional color grading”
- “award-winning photography”
Furthermore, consistent negative prompts help ensure stable output.
Conclusion
The n8n AI Model Workflow gives you a scalable, repeatable, and visual way to automate AI model generation. With just a single click, you can create, process, host, and save high-quality images — all without writing code.
If you plan to expand the workflow, you could add:
- Automatic social media posting
- Additional model variations
- Watermarking
- Background removal
- Multi-image grid generation
I can help you build any of these enhanced workflows as well.