Apple’s new AI-powered image generation tool, Image Playground, has arrived with the latest beta versions of iOS 18.2, iPadOS 18.2, and macOS 15.2. The feature works both as a standalone app and within Messages, though its promised integration with Notes, Freeform, and Pages isn’t available yet.
Getting started
-
After updating your device, look for the Image Playground icon showing a dog or cat in a bubble among your apps.
-
Open the app and tap ‘Request Early Access’ to join the waitlist.
- Access typically takes several days to be granted. While notifications should arrive, you might need to check manually.
The interface is straightforward, featuring a prompt bar at the bottom where you describe your desired image.
- Enter a prompt like ‘otter holding a drink in hands’ and the AI immediately starts generating images.
-
Above the prompt bar, you’ll find suggestion chips for themes, costumes, accessories, and locations to enhance your prompt.
-
Tapping suggestions instantly adds them to your creation, with real-time updates as you stack multiple elements.
- Navigate through generated images using the right arrow or swipe left on iPhone. The on-device processing continues generating variations as you browse.
-
Save images to your library by tapping ‘Done’ in the top-right corner.
-
For more options, tap the three-dot menu below any image to save, copy, or share while continuing to browse.
- Provide feedback using thumbs-up/down buttons. Negative feedback requires additional details.
Image styles
-
The prompt bar area includes options for selecting people, styles, and photos.
-
Currently available styles are Animation and Illustration.
- Animation creates more dimensional 3D-like images, while Illustration produces flat 2D artwork.
Creating personalized images
The standout feature is generating images featuring people from your photo library. All processing happens on-device, ensuring privacy.
- Select ‘People’ at the bottom to see contacts from your Photos app.
- Choose from suggested photos or select your own reference image. This can be modified later.
- Add descriptions or select from suggested scenarios.
- Alternatively, use the Appearance option to specify characteristics like skin tone.
Using reference photos
You can upload any photo as a base image, though faces must be clearly visible for people. The tool works with varying success for animals and places.
Some images might be too ambiguous for the AI to process effectively.
Messages integration
- In a chat, tap the plus icon.
- Select Image Playground from the menu.
- Tap the plus icon to start creating.
- The experience mirrors the main app, with suggestions based on chat context.
- After creating, the image loads directly into the message field.
Current limitations
The tool has several constraints:
- Occasional anatomical errors in generated images
- Strict language requirements with no tolerance for typos
- Limited to animation and illustration styles
- Portrait-only people images
- Difficulty with multiple objects or complex scenes
- Cannot generate text in images effectively
- Restricted content policies
Text generation attempts often produce garbled results:
