I built a bulk AI image generation workflow using Gemini Image API (no Google Cloud setup needed)

Hire a Pro Offer Help
,
1

Hey community :waving_hand:

Sharing a workflow I put together for a client who needed to generate
500+ product images daily using Google’s Gemini Image (Nano Banana) model.

The challenge: Google AI Studio’s rate limits kept killing the workflow
mid-run, and setting up Vertex AI properly was taking days.

The fix: I switched to an OpenAI-compatible Gemini Image endpoint — same
model, same output quality, but the base URL and key just drop into the
HTTP Request node directly. No Google Cloud account, no quota requests.

**The workflow:**

  1. Google Sheets → read product list
  2. HTTP Request → call Gemini Image API (OpenAI format)
  3. Loop over items with 500ms delay between calls
  4. Save output URLs back to Sheets + upload to S3

Works for Nano Banana (fast/cheap) and Nano Banana Pro (4K, better text
rendering) — just swap the model name.

Happy to share the workflow JSON if anyone wants it. Also works cleanly
with Veo 3.1 for video generation using the same endpoint pattern.

Anyone else running image generation workflows at scale in n8n?
Curious what rate limits you’re hitting.

5

This looks great — do you mind sharing what you’re paying per image (especially for Nano Banana vs Pro)? Trying to estimate costs for a similar workflow.