Bulk AI Image Generator: Create Hundreds of Images from a List of Prompts or a Spreadsheet (Nano Banana, GPT Image)

Turn a list of text prompts, or an Excel or CSV file, into a full set of AI images in one run. Pick Google's Nano Banana Pro, Nano Banana 2 and Nano Banana (Gemini image models) or OpenAI GPT Image (GPT-5 Image, GPT-5.4 Image 2), set the aspect ratio and resolution, and get every image back as a downloadable file in a clean table. No API key needed: model usage is billed through your Apify account.

It replaces typing prompts into an image app one at a time and saving each picture by hand.

---

What you can do with it

• Generate images in bulk. Paste 10, 100 or 1,000 prompts, one per line. Each prompt becomes an image.
• Generate from a spreadsheet. Upload an Excel or CSV file of products, articles or ideas and get one image per row, plus your file back with an image link on every row.
• Choose the model. Nano Banana Pro for quality, Nano Banana 2 or Nano Banana 2 Lite for speed and volume, GPT Image for OpenAI's style, or any other image model by id.
• Control the format. Aspect ratios 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3 and 21:9, resolutions from 0.5K up to 4K.
• Make variations. Get up to 10 versions of every prompt and keep the best.
• Keep a consistent style. Add one style instruction (brand colors, illustration style, lighting) and it is applied to every image.
• Download everything. Each image gets a direct link, a thumbnail preview in the Output tab, and full details you can export as Excel, CSV or JSON.

---

Generate images from an Excel or CSV file

Have a product catalog, a list of article titles or a content calendar in a spreadsheet? Upload it and get one image per row.

1. Spreadsheet: upload your .xlsx, .csv or .tsv file, or paste a Google Sheets link shared as "Anyone with the link". The first row must hold the column names.
2. Choose where the prompt comes from:
   • A template: write one prompt with column names in double curly braces in Prompts, for example Studio product photo of a {{Color}} {{Product}} on a plain white background. It is filled in for every row.
   • A prompt column: leave the placeholders out and set Prompt column (or leave it empty: a column named prompt or description is picked automatically, otherwise the column with the longest text).
3. Pick a model, aspect ratio and resolution, and click Start.

What you get back:

• Output tab: one row per image, starting with the spreadsheet row's own columns, with a thumbnail preview.
• Your spreadsheet with image links (Excel): your file in its original row order, with imageUrl, the prompt used and the status added to every row (stored as results.xlsx). With Images per prompt above 1, each row appears once per variation.

row_number	Product	Color	SKU	imageUrl	prompt
1	Ceramic coffee mug	matte black	MUG-01	.../image-00000.png	Studio product photo of a matte black Ceramic coffee mug...
2	Canvas tote bag	olive green	BAG-07	.../image-00001.png	Studio product photo of a olive green Canvas tote bag...

The example prompts prefilled in the form are ignored when a spreadsheet is attached, so you never pay for extra images by accident.

---

Ready-made examples

Open one, click Try it, and the whole setup is filled in for you:

• Product photos for an online store: square catalog shots on a clean background.
• Social media ad creatives: vertical 9:16 images for Stories, Reels and TikTok, 3 variations each.
• Product photos from an Excel or CSV catalog: one studio photo per spreadsheet row, and your catalog back with image links.
• Blog header images from titles: 16:9 hero images in one shared illustration style.

---

Quick start

1. Add your Prompts, one per line.
2. Pick an Image model, an Aspect ratio and an Image resolution.
3. Optional: set Images per prompt for variations, and a System prompt for a shared style.
4. Click Start. When the run finishes, open the Output tab to preview the thumbnails and download the images.

Input example

{
  "prompts": [
    "A photorealistic golden retriever puppy in a sunlit meadow, shallow depth of field",
    "A minimalist flat-design mountain logo inside a circle, blue and white",
    "A cyberpunk city street at night, neon reflections on wet pavement"
  ],
  "model": "google/gemini-3.1-flash-image",
  "systemPrompt": "Cohesive flat-design illustration style, soft pastel palette",
  "aspectRatio": "16:9",
  "imageSize": "2K",
  "imagesPerPrompt": 2,
  "maxItems": 50
}

Output example

One row per image:

{
  "index": 0,
  "prompt": "A photorealistic golden retriever puppy in a sunlit meadow, shallow depth of field",
  "model": "google/gemini-3.1-flash-image",
  "aspectRatio": "16:9",
  "imageSize": "2K",
  "imageUrl": "https://api.apify.com/v2/key-value-stores/XXXX/records/image-00000.png",
  "contentType": "image/png",
  "fileExtension": "png",
  "fileSizeBytes": 1742310,
  "description": "A golden retriever puppy sitting in tall grass...",
  "success": true
}

imageUrl is a direct link to the image file. The Output tab shows it as a thumbnail.

---

Available image models

Model	Best for
Nano Banana Pro (Gemini 3 Pro Image)	Highest quality, detailed scenes, readable text in images
Nano Banana 2 (Gemini 3.1 Flash Image)	Good quality at speed, large batches
Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)	The lightest Nano Banana, for very large volumes
Nano Banana (Gemini 2.5 Flash Image)	Reliable all-rounder, handles short prompts well
GPT-5.4 Image 2, GPT-5 Image, GPT-5 Image Mini	OpenAI's image style and prompt understanding

Need another model? Put any image model id into Custom model ID. The dropdown is refreshed regularly as new image models are released.

Tip: Gemini (Nano Banana) models are the most reliable at always returning an image. If an OpenAI model keeps replying with text instead of a picture, switch to a Gemini model.

---

Use cases

• Ecommerce product photos: catalog shots, lifestyle scenes and mockups for a whole product range.
• Social media and ads: a month of on-brand posts, ad creatives and A/B test variations from a content calendar.
• Blog and SEO images: a unique header image for every article, from the article titles.
• Game and app assets: icons, textures, backgrounds and concept art at scale.
• Design exploration: fan out one idea into dozens of variations and pick the winner.
• Synthetic image datasets: labeled images for machine learning experiments.

---

Why use this instead of an image app

• Bulk by design. Images are generated in parallel, not one click at a time.
• No API keys or code. No Google AI Studio or OpenAI account, no base64 decoding, no rate-limit handling.
• One table for everything. Prompt, model, settings, file link and size for every image, ready to export.
• Reliable long runs. Rate limits and temporary errors are retried automatically, a failed prompt never stops the rest of the run, and a stopped run resumes without generating finished images again.
• Runs in the cloud. Schedule it, call it from the API, or connect it to Zapier, Make or n8n through Apify's integrations.

---

Pricing

This Actor uses pay-per-event pricing: you pay a small fee for each image row the run produces, and the AI model usage of each image is billed through your Apify account. Model usage depends on the model and resolution you pick. The current rates are shown on this Actor's Store page.

• Free Apify plan: runs are limited to 10 images, and AI usage is charged at 10 times the standard token rate. Subscribe to a paid Apify plan for unlimited runs at the standard rate.
• Control your spend: set Max images to cap a run, and set a spending limit on the run in Apify. The Actor stops generating once the limit is reached.
• Failed prompts are reported, not hidden. If a prompt still fails after automatic retries, you get a row with success: false and the reason in error, and that row is billed like any other result. Filter on success to separate failures from images.

{
  "index": 7,
  "prompt": "a portrait of <real person>",
  "model": "google/gemini-3.1-flash-image",
  "imageUrl": null,
  "error": "model returned text, no image (finish_reason=stop): I can't generate images of real people...",
  "success": false
}

---

Frequently asked questions

Do I need an API key for Nano Banana or GPT Image?

No. Model access runs through Apify and is billed to your Apify account, so there is nothing to sign up for. If you prefer to use your own OpenRouter account, paste your key into OpenRouter API key: calls then go straight to OpenRouter and are billed there instead. The per-image fee still applies.

How many images can I generate in one run?

As many as you have prompts, times Images per prompt (up to 10 each). Free Apify plans are limited to 10 images per run. Use Max images to set your own cap.

Can I generate images from a spreadsheet or CSV?

Yes. Upload an Excel, CSV or TSV file (or a Google Sheets link) in Spreadsheet. Each row becomes an image, from a {{Column}} template or from a prompt column, and you get your file back as results.xlsx with an image link on every row.

Why did a prompt come back as text instead of an image?

Some chat-capable image models answer a very short prompt ("flying monkey") with a question instead of drawing. The Actor already tells the model to always return an image, and you can reinforce that in System prompt. Gemini (Nano Banana) models are the most reliable here.

Why did some prompts fail with a timeout when I raised Concurrency?

The built-in AI access starts refusing connections above about 6 parallel images. Keep Concurrency at 4 to 6. With your own OpenRouter key this limit does not apply.

Where are my images stored?

In the run's key-value store, linked from each row's imageUrl. Download them from the links, or export the table as Excel, CSV or JSON.

Which aspect ratios and resolutions are supported?

1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3 and 21:9, at 0.5K, 1K, 2K or 4K. Support varies by model, and a model ignores a setting it does not accept.

Can I use the images commercially?

Usage rights depend on the model provider's terms (Google for Nano Banana, OpenAI for GPT Image). Check them for your use case.

---

Responsible use

You are responsible for your prompts and the images you generate, and for following the model providers' usage policies. Do not generate content that is illegal, infringing or deceptive, or that depicts real people without their consent.

---

More AI Actors by the same author

• Bulk LLM Runner: run ChatGPT, Claude, Gemini and Perplexity prompts in bulk, or on every row of an Excel file.
• Bulk AI Translator: translate Excel and CSV files, websites, PDFs and subtitles into 40+ languages.
• Bulk Text to Speech: turn texts or a spreadsheet into MP3 voiceovers with subtitles in 75 languages.
• AI Brand Visibility Tracker: find out whether ChatGPT, Perplexity, Gemini, Claude and Grok recommend your brand.

💼 Need a custom workflow? Reach out at youssefarhan24@gmail.com