Krea 3
Krea 3 is the newest image model from Krea. It generates stills from a text prompt at 1K or 2K, in 1:1, 4:3, 3:4, 16:9, and 9:16, and edits or combines up to 8 reference images while keeping the details you name. Refer to attached images in the prompt as `<image-1>`, `<image-2>`, and so on. Use it for product restaging, character consistency, and material or color changes.
See more
Krea 3 is the newest image model from Krea. It generates stills from a text prompt at 1K or 2K, in 1:1, 4:3, 3:4, 16:9, and 9:16, and edits or combines up to 8 reference images while keeping the details you name. Refer to attached images in the prompt as `<image-1>`, `<image-2>`, and so on. Use it for product restaging, character consistency, and material or color changes.
Connect Claude, Cursor, VS Code, or Codex to Layer’s MCP server and Krea 3 becomes a tool your agent can reach for — it picks the parameters, runs the generation, and shows you the result in the conversation.
https://mcp.app.layer.ai/mcpAsk your agent
“Use Krea 3 to generate 4 variations of a crystal sword icon on a transparent background.”
execute_forge({
"base_model_id": "krea-3",
"prompt": "A crystal sword icon on a transparent background, game UI art",
"batch_size": 4,
"width": 1024,
"height": 1024
})Authenticate with a personal access token, post the model’s own form, and poll the inference until it completes. Every field below is accepted by this endpoint.
# Start the generation — returns an inference id straight away.
curl -X POST https://api.app.layer.ai/api/v2/workspaces/$WORKSPACE_ID/base-models/krea-krea-3/inferences \
-H "Authorization: Bearer $LAYER_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"prompt": "A crystal sword icon on a transparent background, game UI art"
}'
# Poll until status is "complete" — the response then carries the asset URLs.
curl https://api.app.layer.ai/api/v2/workspaces/$WORKSPACE_ID/inferences/$INFERENCE_ID \
-H "Authorization: Bearer $LAYER_TOKEN"- prompt stringrequired
- Describe what to generate, or the change to make.
- negative_prompt string
- What to keep out of the result.
- aspect_ratio enumdefault SQUARE_1K13 options
- The shape and size of the output. Auto follows the shape of the input.
- seed integer
- The starting point for the model's randomness. The same seed and settings give the same result.
- guidance_file_init_image file[]
- The image to modify
- guidance_file_reference_image file[]
- Guides the edit
- output_fit any
- Trim the transparent margin, and place the image on a canvas of a fixed size.
- prompt_language string
- The language the prompt is written in, so it reaches the model as you meant it.
Why use Krea 3 for image?
High-quality image generation
Krea 3 delivers production-ready outputs with clean structure, sharp detail, and strong prompt adherence. Ideal for character art, environments, and marketing assets.
Strong prompt control
The model interprets creative instructions with high accuracy, even when switching between different styles, lighting setups, or character directions.
Clean compositions for UA
Maintains stable framing, clear silhouettes, and high focal clarity. Assets read well at mobile sizes, making it strong for UA experiments and ad creatives.
Rapid character and cosmetic ideation with strong structure
Quickly explore character poses, cosmetic ideas, and early art directions. Clean silhouettes and expressive results help teams evaluate concepts before moving to heavier models.
Fast UA creative exploration and thumbnail testing
Generate clear, high-readability frames that work well for quick A/B testing across networks. Test multiple creative angles and iterate quickly based on results.
Lightweight environment and item exploration for LiveOps
Draft environments, themed content, shop items, and event visuals in a rapid loop. Ideal for early-stage direction setting before committing to final polish.
FAQ
How much does Krea 3 cost? Is it free?+
How does Layer's pricing work?+
There are no seat fees, feature gates, or plans on Layer. Instead, our platform uses a consumption based system with Creative Units (CUs) with a flexible monthly subscription. Every generation on Layer (image, video, 3D, or audio) consumes a Creative Unit, and you only pay for what you create.
Other image models on Layer
Krea 2 Large
Krea
Krea's flagship text-to-image model for high-fidelity generations with distinctive aesthetic range.
Krea 2 Turbo
Krea
Speed-optimized open-source version of Krea 2 — high-fidelity images in seconds, with the full Krea aesthetic range.
FLUX.3 Klein Base Edit
Black Forest Labs
Black Forest Labs' FLUX.3 Klein Base editor for multi-reference edits at 1K to 4K.
Kling Image O3
Kling
Kling Omni 3 image model. High-quality images with text rendering capabilities up to 4K resolution.
FLUX.3 Klein Base
Black Forest Labs
Black Forest Labs' FLUX.3 Klein Base for native 1K to 4K text-to-image.
Recraft Creative Upscale
Recraft
Upscale for a sharper, cleaner, higher-resolution result.