Skip to content

Latest commit

 

History

History
104 lines (74 loc) · 3.32 KB

File metadata and controls

104 lines (74 loc) · 3.32 KB

Google Play feature graphics

← Back to the README · CLI reference · Configuration

SnapAI generates Google Play feature graphics at exactly 1024x500. It creates the artwork with OpenAI or Gemini, then resizes it to the required store dimensions.

Use either the full command or its fg alias:

npx snapai feature-graphic --prompt "fitness tracker app"
npx snapai fg --prompt "fitness tracker app"

Add an app name

Use --app-name to ask the model to render the product name as part of the banner:

npx snapai fg \
  --prompt "energetic fitness tracker with bold green gradients" \
  --app-name "FitTrack"

Image models can still misspell text. Inspect the result before uploading it to Google Play.

Composite an existing logo

Use --logo when you already have a logo and want SnapAI to place it on the generated banner:

npx snapai fg \
  --prompt "clean productivity banner with a blue-to-purple gradient" \
  --logo ./assets/icon.png \
  --logo-position left

--logo-position accepts left, center, or right.

Choose a model

Feature graphics support:

  • gpt-2 using OpenAI gpt-image-2
  • gpt-1.5 using OpenAI gpt-image-1.5
  • gpt-1 using OpenAI gpt-image-1
  • banana using Gemini gemini-2.5-flash-image
  • banana-2 using Gemini gemini-3.1-flash-image-preview
  • banana --pro using Gemini gemini-3-pro-image-preview

gpt-2 is the default and SnapAI's friendly alias for gpt-image-2. The official gpt-image-2 ID and gpt-image-2-2026-04-21 snapshot are also accepted. SnapAI requests a native 1280x640 GPT Image 2 canvas before cropping and resizing it to the required 1024x500 output.

npx snapai fg \
  --prompt "weather app with dramatic clouds and clear negative space" \
  --model gpt-2

Format and variations

Google Play feature graphics can be written as PNG or JPEG. WebP is intentionally unavailable for this command.

npx snapai fg \
  --prompt "travel planner with a warm editorial style" \
  --model gpt-2 \
  --output-format jpeg \
  --output-compression 85

--output-compression 0-100 is available for OpenAI JPEG output. It is not supported with PNG.

OpenAI models and Banana Pro support up to 10 variations through -n. Regular banana and banana-2 support one image.

npx snapai fg \
  --prompt "music player with an energetic neon sound wave" \
  --model gpt-1.5 \
  -n 3

Prompt controls

Feature graphics support the same core prompt controls as icon generation:

  • --style adds a built-in or custom visual direction.
  • --raw-prompt disables SnapAI’s prompt enhancement.
  • --prompt-only previews the final prompt without generating an image.
npx snapai fg \
  --prompt "calculator app" \
  --style minimalism \
  --prompt-only

Design tips

  • Keep the core subject away from the edges so resizing does not make the composition feel cramped.
  • Leave deliberate negative space for a logo or app name.
  • Use a short app name if asking the model to render text.
  • Prefer --logo for an existing brand mark instead of asking the model to recreate it.
  • Preview the enhanced prompt with --prompt-only before generating expensive variations.

For every option, see the snapai feature-graphic flag reference.