Image Generation

Nano Banana

Image generation by Google. Supports up to 14 reference images, web grounding, text rendering, and output up to 4K.

SyncAll plans

Nano Banana 2

Best speed + quality, text rendering, up to 14 refs, web grounding

nano-banana-2

Nano Banana Pro

Best for text and multilingual content

nano-banana-pro

Generate

POST/api/v1/generate

Image generation is synchronous -- the response contains the output URL(s) immediately. No polling needed. Credits are deducted upfront and auto-refunded if generation fails.

Parameters

model

stringrequired

nano-banana-2 or nano-banana-pro

prompt

stringrequired

Text prompt describing the image. Supports @image1 through @image10 to reference attached images.

aspect_ratio

string

Nano Banana 2: 1:1 16:9 9:16 4:3 3:4 3:2 2:3 4:5 5:4 21:9 1:4 4:1 1:8 8:1
Default: 1:1

resolution

string

512 1K 2K 4K. Default: 1K. NB Pro defaults to 2K.

reference_images

string[]

Up to 10 public image URLs. Use @image1, @image2 in prompt. Upload via /api/v1/upload.

grounding

boolean

Enable web grounding via Google Search. Improves accuracy for real-world entities, logos, landmarks, recent events. Default: false

tag

string

Tag the output for reuse: character object style location. Tagged outputs appear in your library.

Reference Images

Attach up to 10 images as visual references. Use @image1, @image2, etc. in your prompt to direct each reference. All URLs must be publicly accessible -- use /api/v1/upload to host files.

Common use cases:

Character consistency -- keep the same character across multiple images
Style transfer -- match the visual style of a reference image
Product shots -- place a product (from ref) in different scenes
Edit / remix -- modify an existing image with a text prompt
Multi-subject -- combine elements from different references

Web Grounding

When grounding: true, the model uses Google Search to ground its knowledge in real-world data. This dramatically improves accuracy for:

Real people -- accurate likenesses of public figures
Logos & brands -- correct brand marks and colors
Landmarks -- accurate depictions of real places
Recent events -- knowledge beyond the model's training cutoff
Technical diagrams -- factually accurate illustrations

Web grounding adds a small per-request surcharge to the credit cost.

Examples

Basic image generation

curl -X POST https://www.lunostudio.ai/api/v1/generate \
  -H "Authorization: Bearer luno_sk_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-2",
    "prompt": "A cozy coffee shop interior with warm lighting, watercolor style",
    "aspect_ratio": "16:9",
    "resolution": "2K"
  }'

Character consistency with refs

curl -X POST https://www.lunostudio.ai/api/v1/generate \
  -H "Authorization: Bearer luno_sk_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-2",
    "prompt": "@image1 sitting at a desk working on a laptop, same outfit and hairstyle",
    "reference_images": ["https://storage.lunostudio.ai/.../character.jpg"],
    "resolution": "2K",
    "tag": "character"
  }'

Multi-ref composition

curl -X POST https://www.lunostudio.ai/api/v1/generate \
  -H "Authorization: Bearer luno_sk_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-2",
    "prompt": "@image1 wearing the outfit from @image2, standing in the location from @image3",
    "reference_images": [
      "https://storage.lunostudio.ai/.../person.jpg",
      "https://storage.lunostudio.ai/.../outfit.jpg",
      "https://storage.lunostudio.ai/.../location.jpg"
    ]
  }'

Web grounding for real-world accuracy

curl -X POST https://www.lunostudio.ai/api/v1/generate \
  -H "Authorization: Bearer luno_sk_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-2",
    "prompt": "The Eiffel Tower at sunset with cherry blossoms in the foreground, photorealistic",
    "grounding": true,
    "resolution": "4K",
    "aspect_ratio": "9:16"
  }'

Response (immediate)

{
  "output": ["https://storage.lunostudio.ai/generations/user/api-1716234567.png"],
  "credit_cost": 3,
  "model": "google/nano-banana-2",
  "type": "image"
}

Tagging Outputs

Pass tag to categorize the output for reuse in future generations:

characterA person or character to maintain consistency across generations

objectA product, item, or prop to place in different scenes

styleA visual style or aesthetic to apply to other prompts

locationA place or environment to reuse as a backdrop

Tagged outputs are saved to your library and visible in your dashboard at /requests.

Seedance 2.0 Files API Back to Overview

Need help?

Engineering team responds within 5 minutes.

Discord support@lunostudio.ai matt.sharpe@lunostudio.ai