🤯 This INSANE Nano Banana + n8n Workflow Just DESTROYED Traditional Product Photography (Generated 500+ Photos in 1 Hour!)

Just finished building a complex workflow using Google's new Nano Banana AI model through Fal.ai. Creates professional product photos by combining any model image with t-shirt designs - completely automated from Excel input to WooCommerce upload.

How it Works (Step-by-Step):

1. Data Input & Batching:

Manual trigger reads Excel file (model images + t-shirt designs + product IDs)
Split into batches of 10 to prevent API overload
Each row gets processed with model + design combination

2. Smart Caching System:

Code node generates unique cache key for each image combo
Checks if combination was processed before
If cache exists → skip API call, use stored result
If no cache → proceed to image generation

3. AI Image Generation:

HTTP request to Fal.ai using Nano Banana model
Prompt: "Photo of model wearing submitted clothing item, professional product photography"
Smart Wait node polls every 5-10 seconds for completion (prevents rate limits)

4. Status Checking & Error Handling:

Check Status node verifies successful generation
If failed → loops back to Smart Wait
If successful → proceeds to download

5. Storage & Integration:

Downloads generated images
Uploads to Google Drive for storage
Updates Excel with new image URLs
Optional: Auto-pushes to WooCommerce/Shopify

Key Technical Features:

Batch processing prevents API failures
Caching system eliminates duplicate charges
Error handling with retry loops
Rate limiting with smart wait intervals

Example Output: Takes a model photo + t-shirt design → generates professional product photo showing model wearing the shirt

The caching is the game-changer here - once an image combo is generated, it never gets charged again. Client processes 50+ products monthly and only pays for new combinations.

Built this over 2 weeks of testing different approaches. The Smart Wait node timing was crucial - too fast and Fal.ai rejects requests, too slow and workflow takes forever.

3 comments