Image Generation

Krea 3 API

Krea 3 for reference-driven images and edits. Start from up to 8 images, describe the change and keep the parts you like, or generate from a prompt alone. Up to 4K in 10 aspect ratios. Integrate this model via REST with endpoint docs, parameters, and code examples.

Prompting guide Compare image models Use cases

Img2Img

Editing

Credit Cost

1.5 credits

Type

Image

Get API Key

Create auth keys for apps and agents

See Usage

Track requests, credits, and logs

Buy Credits

Top up prepaid API capacity

API Endpoint

POST

/api/v1/models/krea-3/run

Test This Model

Run a real request through the API pipeline and inspect the exact input and output your integration will handle. Uses your credits like any API call; failed runs are refunded automatically.

LLM.txt

AI-Ready

For AI assistants

Copy this model's documentation into Claude, GPT, or any AI for instant integration help.

/llm/krea-3.txt

OpenAPI Spec

Swagger compatible

Import into Postman, Swagger UI, or generate client SDKs in any language.

/api/openapi

Parameters

Input parameters accepted by this model

Parameter	Type	Description	Default
`prompt` required	string	What to make, or what to change when image_urls is set. Refer to references by position ("the jacket from the second image") and to edit regions as @label or @1, @2.	-
`image_urls`	array	Optional, up to 8. The first image is the one being edited; the others bring in objects, people or style. Each reference costs about as much as the output again. Four or fewer keep full detail. items: url min items: 1 \| max items: 8	-
`edit_regions`	array	Optional, up to 24 boxes, needs image_urls. Each is { reference_index, x, y, width, height } with coordinates 0 to 1 of that image, plus an optional label (no spaces, call it as @label) and note (up to 600 characters, overrides the prompt inside the box). Boxes on the first image mark where the edit happens; boxes on the others mark objects to bring in. items: object max items: 24	-
`thinking_level`	enum	fast (default), or high for a stronger planning step on multi-step edits and many references. High takes longer. `fasthigh`	`fast`
`resolution`	enum	Output size: 1K (about 1 MP), 2K (about 4 MP) or 4K (about 16 MP). `1K2K4K`	`2K`
`aspect_ratio`	enum	Output shape. Defaults to 1:1 without references. With references, leave it out to keep the input shape, or set it to reframe. `1:14:53:42:39:16`+5 more	-
`num_images`	integer	Number of images to generate (1-4) min: 1 \| max: 4	`1`
`seed`	integer	Integer seed for reproducible outputs.	-
`output_format`	enum	Image output format `pngjpegwebp`	`png`

Use saved characters (@Name)

Write @Name in the prompt to use a character, product, outfit, location, prop or style saved in the account's Character Studio, for example "@Mara on a beach at sunset". Its photos are added to image_urls automatically, after any images you send yourself, so your own image numbering is unchanged. List saved names with GET /api/v1/characters and save new ones with POST /api/v1/characters. An @word that names nothing saved stays plain text.

Supported Aspect Ratios

1:1

4:5

3:4

2:3

9:16

5:4

4:3

3:2

16:9

21:9

Code Examples

Text to image: copy and paste to get started

bash

curl -X POST "https://pixeldojo.ai/api/v1/models/krea-3/run" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "a quiet canal town at golden hour, film photograph",
    "resolution": "2K",
    "aspect_ratio": "3:2"
  }'

Response Format

Submit Response (202 Accepted)

json

{
  "jobId": "job_abc123...",
  "status": "pending",
  "statusUrl": "https://pixeldojo.ai/api/v1/jobs/job_abc123...",
  "creditCost": 1.5,
  "creditsRemaining": 95,
  "expiresAt": "2026-11-06T14:03:23Z"
}

Completed Response