YARSY AI Studio — Image Generation

🍌

YARSY AI Studio

—

Main prompt

Prompt variations Each variation generates a separate image

Add reference image (0/5)

Generated image will appear here

~$0.04 per image

Model

Choose AI model for generation. Nano Banana models support native image gen + editing. Imagen is photo-focused.

View all models →

System instructions

Set tone, style, or rules for the model. Example: "Generate in minimalist style with pastel colors."

Learn more →

Add new instruction

Temperature

Controls randomness. Lower = more predictable, higher = more creative/varied.

• 0.0-0.5: Consistent results
• 0.7-1.0: Balanced
• 1.0-2.0: Very creative

1.0

Aspect ratio

Output image dimensions.

• 1:1 — Square (Instagram)
• 16:9 — Landscape (YouTube)
• 9:16 — Portrait (Stories)
• 4:3 — Classic photo

Auto

Model decides

1:1

Square

9:16

Portrait / Stories

16:9

Landscape / YouTube

3:4

Portrait photo

4:3

Classic photo

3:2

DSLR standard

2:3

Portrait DSLR

5:4

Large format

4:5

Instagram portrait

21:9

Ultrawide / Cinema

Resolution

Output image resolution.

• 1K — 1024px (fast, drafts)
• 2K — 2048px (balanced)
• 4K — 4096px (print quality)

Note: 4K only with Nano Banana 2

Number of images

Generate multiple variations at once. Each can have a unique prompt variation.

Cost multiplies by number of images.

Advanced settings

Output length

Max tokens in model response. Higher = longer text descriptions with images.

32768

Top P

Nucleus sampling. Controls diversity by limiting token pool. Lower = more focused, higher = more varied.

Learn more →

0.95

Stop sequence

Text that stops generation when encountered. Useful for structured outputs.

Prompting guide & docs

Welcome back