🎨 AI Image Generator — Text to Image with 7 Models

Turn any text prompt into stunning images. Pick from 7 AI models, 10 aspect ratios, and generate up to 4 images at once — all for a flat $4 per 1,000 images.

No API keys. No sign-up. Just describe what you want and get high-quality images in seconds.

---

✨ Why this Actor

• 🤖 7 models, one price — switch styles freely, every model costs the same
• 🖌️ Text-to-image AND image editing — generate from scratch or edit an existing photo
• 📐 10 aspect ratios — square, portrait, landscape, ultrawide
• ⚡ Parallel generation — up to 4 images at once, not one-by-one
• 🔁 Never fails silently — if a model is busy, it auto-switches to a working one
• 🔗 Permanent links — every image is re-hosted on a stable CDN
• 🔒 Zero setup — no keys, no login, no config

---

🖼️ Models

Model	Style / Strength
Nano Banana 2 Lite ⭐	Sharp details, great all-rounder (recommended)
Nano Banana	Consistent subjects, clean edits
Gempix 2	Vibrant colors, bold creative looks
Seedream 4	Reliable prompt following
GPT Image 1.5	Versatile, clear compositions
Qwen Image	Strong text rendering & layouts
Minimax Image 01	Fast drafts and quick concepts

All models are billed at the same flat rate — experiment freely.

---

🚀 Quick Start

{
  "prompt": "A futuristic city at sunset, neon lights, cinematic, ultra detailed",
  "model": "nano-2-lite",
  "aspectRatio": "16:9",
  "resolution": "2K",
  "numImages": "2"
}

Run it → get 2 ready-to-use image URLs in the dataset.

---

🖌️ Editing an Existing Photo (image-to-image)

Provide imageUrl to edit a photo instead of generating from scratch. The prompt then describes the change you want.

{
  "prompt": "add a colorful party hat on the cat and a festive background",
  "imageUrl": "https://example.com/cat.jpg",
  "model": "nano-banana"
}

Best editing models: Nano Banana, Qwen Image, Seedream 4. Leave imageUrl empty for normal text-to-image.

---

⚙️ Input

Field	Type	Required	Default	Description
prompt	string	✅	—	What you want to see (or the change to make when editing)
imageUrl	string	❌	—	URL of a photo to edit (image-to-image). Empty = text-to-image
model	string	❌	nano-2-lite	Which AI model to use
aspectRatio	string	❌	1:1	Image shape
resolution	string	❌	1K	Output resolution: 1K, 2K, 4K
numImages	string	❌	1	How many to generate (1, 2, or 4)
timeout	int	❌	120	Max seconds to wait per image

Resolution

Level	Detail	Speed
1K	Standard	Fastest
2K	Higher detail	Medium
4K	Maximum detail	Slower

Not every model supports every resolution. If you request a resolution a model can't produce, it delivers the closest available one — the output always reports the actual resolution plus exact width/height, so you know precisely what you got.

Model	Supported resolutions
Nano Banana 2 Lite	1K, 2K, 4K
Nano Banana	1K, 2K, 4K
Gempix 2	1K, 2K, 4K
Seedream 4	1K, 2K, 4K
GPT Image 1.5	1K, 2K
Qwen Image	1K, 2K
Minimax Image 01	1K only

Same price at every resolution.

Aspect ratios

Ratio	Perfect for
1:1	Social posts, avatars, product tiles
16:9	Thumbnails, banners, slides
9:16	Stories, Reels, phone wallpapers
4:3 / 3:4	Blog headers, Pinterest pins
3:2 / 2:3	Classic photography framing
21:9	Cinematic ultrawide

---

📤 Output

Each image becomes one dataset item:

{
  "success": true,
  "prompt": "A futuristic city at sunset, neon lights, cinematic, ultra detailed",
  "model": "nano-2-lite",
  "aspectRatio": "16:9",
  "resultUrl": "https://api.apify.com/v2/key-value-stores/.../records/gen-abc123.jpeg",
  "format": "jpeg"
}

If your chosen model was busy and a fallback was used, you'll also see:

{
  "requestedModel": "gpt-image-1-5",
  "note": "requested model was busy; used a fallback model"
}

---

💰 Pricing

	Price
Per image	$0.004
1,000 images	$4

Simple and flat — every model, every size, same price. You only pay for images that generate successfully.

---

💡 Use Cases

• 🎨 Social media — fresh, unique visuals for every post
• 🛍️ E-commerce — product mockups and lifestyle shots
• 📰 Blogs & articles — custom headers and featured images
• 📱 App & web design — UI illustrations and placeholders
• 🎮 Game & concept art — characters, scenes, moodboards
• 📊 Presentations — bespoke slides and diagrams

---

❓ FAQ

Do I need an account or API key? No. Just enter a prompt and run.

What if a model is overloaded? The Actor automatically retries and switches to a reliable model so you still get your images.

Can I generate multiple images at once? Yes — set numImages to 2 or 4. They generate in parallel.

Which model should I pick? Start with Nano Banana 2 Lite (default). For vibrant art try Gempix 2; for fast drafts try Minimax.

Can I edit my own photo? Yes. Put its URL in imageUrl and describe the change in prompt. Works best with Nano Banana or Qwen Image.