Krea 3 Image Edit API

Krea 3 Image Edit (Image-to-Image)

What is Krea 3 Image Edit?

Krea 3 Image Edit is Krea's editing-first foundation image model. You give it one to eight reference images and a plain-language instruction — say what to change and what to keep — and it returns the edit at native 2K. Its defining strength is restraint: it leaves alone everything you did not ask it to touch, so a product's shape, label and logo stay pixel-exact while a brand-new scene is built around it. The same model restyles, recomposes, swaps products and materials, removes objects, and combines several numbered references into one coherent picture.

Key Features

Product and label preservation — change the scene around a product and its label, logo and form stay sharp, turning a plain packshot into a campaign photo.
Instruction-following edits — write edits in everyday language: lead with the change, close with one preservation sentence.
Up to eight references — point it at a person, a product, a place and a style, number them, and it composes one image.
Regional edit — confine a change to one region and keep the rest pixel-exact, so edits chain without quality loss.
Native 2K output with a lighter 1K option, plus fast and high thinking modes.

Best Use Cases

Krea 3 Image Edit suits teams that edit at volume: e-commerce and marketing turning one packshot into lifestyle or on-model shots, campaign variants in many colourways or formats, product and object swaps, object removal and photo cleanup, and keeping the same character across a series. Regional editing makes it safe to update one approved detail without re-rendering the whole frame.

Prompt Tips and Output Quality

Lead with the change, add only the details that help execute it, then close with one sentence naming what must stay (the label, the logo, the person). When using more than one reference, refer to each by what it shows. In testing, a plain white candle-tin packshot became a fully styled scene while the printed label stayed identical. Send four or fewer references when fine detail matters.

How to Use Krea 3 Image Edit

Krea 3 Image Edit transforms images you already have. Upload one to eight reference images, then describe the edit in plain language: say what should change and, in one closing sentence, what must stay the same. The model is tuned to leave untouched anything you do not mention, which is why products, labels and faces survive big changes to the scene around them.

Writing the prompt

Lead with the change and add only the details that help execute it — the new scene, the light, the camera angle. Close with one preservation sentence such as "Keep the product exactly as it is, including the label." Avoid long lists of what to keep; naming the one or two things that matter works better. When a prompt uses more than one reference, refer to each by what it shows, for example "put the jacket from the second image on the person in the first."

Choosing settings

Set resolution to 2k for final, delivery-quality edits and 1k for fast drafts. Leave thinking_mode on fast for clear, everyday edits; switch to high for complex, multi-step or multi-reference edits, or when a fast result missed the request. Turn on regional_edit for local changes — recolour, swap, add, remove or fix one element — so the rest of the frame stays pixel-exact and you can chain edits without quality loss. Leave it off for whole-scene rebuilds, restyles or combining several images. Use aspect_ratio auto to follow the reference framing, or pick a fixed ratio to reframe.

Recommended Settings

Product scene build (tested): resolution 2k, thinking_mode fast, regional_edit false — a plain white packshot became a styled campaign scene in about 20 seconds with the printed label kept identical.
Local fix or recolour: regional_edit true, thinking_mode fast.
Complex multi-reference composite: thinking_mode high, with four or fewer full-detail references.

Examples to try

Each example ran end to end at the settings shown — native 2K, aspect ratio on auto — keeping the detail its closing sentence names while the rest is rebuilt.

Product scene build. Prompt: "Place the candle tin on a weathered zinc tabletop beside a folded charcoal wool throw and an open box of matches, with a limewashed plaster wall behind it. Keep the tin and its printed EMBER label exactly as they are." The white packshot becomes a styled surface; the tin and its EMBER wordmark stay identical. Close with the one detail to keep.
Regional precision edit (regional_edit true). Prompt: "Change only the teapot to glossy emerald green, keeping the teacup, the croissants, the jar of jam, the napkin and the oak table exactly as they are." Only the teapot changes colour; everything else stays pixel-exact, so local edits chain cleanly.
Sketch to a finished render (thinking_mode high). Prompt: "Render this marker sketch of the cordless drill as a photorealistic studio product shot — matte charcoal body, orange rubber grip, brushed-metal chuck — on a seamless grey background. Keep the drill's proportions, the trigger and the battery pack exactly as drawn." The flat drawing becomes a photographed tool in the materials you name; let high mode reason from the lines.

Reference-guided composites

Krea 3 takes up to eight references in one edit — number them and say what each shows, four or fewer when detail matters.

Multi-reference composite (three references, thinking_mode high). Prompt: "Furnish the empty room in image 1 with the tan leather chair from image 2 and the brass arc lamp from image 3, placing the chair by the window and the lamp arcing over it. Keep the room's oak floor, white walls and window exactly as they are." The exact chair and lamp land in the referenced room, lit to the window.

Try instructions like "Remove only the parked car on the left", "Reproduce the same character in a new pose", or "Restyle this photo as a pen-and-ink drawing" — suggestions to adapt, not runs we shipped.

Krea 3 Image Edit examples

Requests run on Krea 3 Image Edit in the playground: what each was given, what came back, and the API call that makes it.

How to restage a candle product photo in a lifestyle scene with Krea 3 Image Edit

Prompt

Place the candle tin on a weathered zinc tabletop beside a folded charcoal wool throw and an open box of matches, with a limewashed plaster wall behind it. Keep the tin and its printed EMBER label exactly as they are.

Input image

Settings

resolution: 2k
aspect_ratio: auto
thinking_mode: fast
regional_edit: false

Place the candle tin on a weathered zinc tabletop beside a folded charcoal wool throw and an open box of matches, with a limewashed plaster wall behind it. Kee… — Output from Krea 3 Image Edit

API request

# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/krea-3-image-edit" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "images": [
    "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-input.jpg"
  ],
  "prompt": "Place the candle tin on a weathered zinc tabletop beside a folded charcoal wool throw and an open box of matches, with a limewashed plaster wall behind it. Keep the tin and its printed EMBER label exactly as they are.",
  "resolution": "2k",
  "aspect_ratio": "auto",
  "thinking_mode": "fast",
  "regional_edit": false
}' | jq -r .request_id)

# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID/status" \
  -H "x-api-key: YOUR_API_KEY"

# 3. Fetch the result
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID" \
  -H "x-api-key: YOUR_API_KEY"

# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind

result = segmind.run(
    "krea-3-image-edit",
    images=[
        "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-input.jpg",
    ],
    prompt="Place the candle tin on a weathered zinc tabletop beside a folded charcoal wool throw and an open box of matches, with a limewashed plaster wall behind it. Keep the tin and its printed EMBER label exactly as they are.",
    resolution="2k",
    aspect_ratio="auto",
    thinking_mode="fast",
    regional_edit=False,
)
print(result.get("output"))

How to recolor one object in a breakfast photo with Krea 3 Image Edit

Prompt

Change only the teapot to glossy emerald green, keeping the teacup, the croissants, the jar of jam, the napkin and the oak table exactly as they are.

Input image

Settings

resolution: 2k
aspect_ratio: auto
thinking_mode: fast
regional_edit: true

Change only the teapot to glossy emerald green, keeping the teacup, the croissants, the jar of jam, the napkin and the oak table exactly as they are. — Output from Krea 3 Image Edit

API request

# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/krea-3-image-edit" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "images": [
    "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-regional-edit-29830f35.jpg"
  ],
  "prompt": "Change only the teapot to glossy emerald green, keeping the teacup, the croissants, the jar of jam, the napkin and the oak table exactly as they are.",
  "resolution": "2k",
  "aspect_ratio": "auto",
  "thinking_mode": "fast",
  "regional_edit": true
}' | jq -r .request_id)

# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID/status" \
  -H "x-api-key: YOUR_API_KEY"

# 3. Fetch the result
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID" \
  -H "x-api-key: YOUR_API_KEY"

# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind

result = segmind.run(
    "krea-3-image-edit",
    images=[
        "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-regional-edit-29830f35.jpg",
    ],
    prompt="Change only the teapot to glossy emerald green, keeping the teacup, the croissants, the jar of jam, the napkin and the oak table exactly as they are.",
    resolution="2k",
    aspect_ratio="auto",
    thinking_mode="fast",
    regional_edit=True,
)
print(result.get("output"))

How to virtually furnish an empty room photo from product shots with Krea 3 Image Edit

Prompt

Furnish the empty room in image 1 with the tan leather chair from image 2 and the brass arc lamp from image 3, placing the chair by the window and the lamp arcing over it. Keep the room's oak floor, white walls and window exactly as they are.

Input images

Settings

resolution: 2k
aspect_ratio: auto
thinking_mode: high
regional_edit: false

Furnish the empty room in image 1 with the tan leather chair from image 2 and the brass arc lamp from image 3, placing the chair by the window and the lamp arc… — Output from Krea 3 Image Edit

API request

# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/krea-3-image-edit" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "images": [
    "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-multi-reference-e8af731a.jpg",
    "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-multi-reference-88ec6ad2.jpg",
    "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-multi-reference-e3919b76.jpg"
  ],
  "prompt": "Furnish the empty room in image 1 with the tan leather chair from image 2 and the brass arc lamp from image 3, placing the chair by the window and the lamp arcing over it. Keep the room'\''s oak floor, white walls and window exactly as they are.",
  "resolution": "2k",
  "aspect_ratio": "auto",
  "thinking_mode": "high",
  "regional_edit": false
}' | jq -r .request_id)

# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID/status" \
  -H "x-api-key: YOUR_API_KEY"

# 3. Fetch the result
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID" \
  -H "x-api-key: YOUR_API_KEY"

# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind

result = segmind.run(
    "krea-3-image-edit",
    images=[
        "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-multi-reference-e8af731a.jpg",
        "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-multi-reference-88ec6ad2.jpg",
        "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-multi-reference-e3919b76.jpg",
    ],
    prompt="Furnish the empty room in image 1 with the tan leather chair from image 2 and the brass arc lamp from image 3, placing the chair by the window and the lamp arcing over it. Keep the room's oak floor, white walls and window exactly as they are.",
    resolution="2k",
    aspect_ratio="auto",
    thinking_mode="high",
    regional_edit=False,
)
print(result.get("output"))

How to turn a product sketch into a studio product shot with Krea 3 Image Edit

Prompt

Render this marker sketch of the cordless drill as a photorealistic studio product shot — matte charcoal body, orange rubber grip, brushed-metal chuck — on a seamless grey background. Keep the drill's proportions, the trigger and the battery pack exactly as drawn.

Input image

Settings

resolution: 2k
aspect_ratio: auto
thinking_mode: high
regional_edit: false

Render this marker sketch of the cordless drill as a photorealistic studio product shot — matte charcoal body, orange rubber grip, brushed-metal chuck — on a s… — Output from Krea 3 Image Edit

API request

# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/krea-3-image-edit" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "images": [
    "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-sketch-render-31f1b79d.jpg"
  ],
  "prompt": "Render this marker sketch of the cordless drill as a photorealistic studio product shot — matte charcoal body, orange rubber grip, brushed-metal chuck — on a seamless grey background. Keep the drill'\''s proportions, the trigger and the battery pack exactly as drawn.",
  "resolution": "2k",
  "aspect_ratio": "auto",
  "thinking_mode": "high",
  "regional_edit": false
}' | jq -r .request_id)

# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID/status" \
  -H "x-api-key: YOUR_API_KEY"

# 3. Fetch the result
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID" \
  -H "x-api-key: YOUR_API_KEY"

# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind

result = segmind.run(
    "krea-3-image-edit",
    images=[
        "https://segmind-resources.s3.amazonaws.com/input/krea-3-image-edit-example-sketch-render-31f1b79d.jpg",
    ],
    prompt="Render this marker sketch of the cordless drill as a photorealistic studio product shot — matte charcoal body, orange rubber grip, brushed-metal chuck — on a seamless grey background. Keep the drill's proportions, the trigger and the battery pack exactly as drawn.",
    resolution="2k",
    aspect_ratio="auto",
    thinking_mode="high",
    regional_edit=False,
)
print(result.get("output"))

Does it keep labels and logos intact?

It preserves products, labels and faces well, but check small print before production.

How many reference images can I use?

Up to eight; send four or fewer when fine detail matters.

What is regional edit?

It regenerates only the region you describe and leaves the rest pixel-exact.

What resolution does it output?

Native 2K (about 4 MP) or a lighter 1K (about 1 MP).

Krea 3 Image Edit

Krea 3 Image Edit (Image-to-Image)

What is Krea 3 Image Edit?

Key Features

Best Use Cases

Prompt Tips and Output Quality

How to Use Krea 3 Image Edit

Writing the prompt

Choosing settings

Recommended Settings

Examples to try

Reference-guided composites

Krea 3 Image Edit examples

How to restage a candle product photo in a lifestyle scene with Krea 3 Image Edit

How to recolor one object in a breakfast photo with Krea 3 Image Edit

How to virtually furnish an empty room photo from product shots with Krea 3 Image Edit

How to turn a product sketch into a studio product shot with Krea 3 Image Edit

Does it keep labels and logos intact?

How many reference images can I use?

What is regional edit?

What resolution does it output?

Related Models

Krea 3 Text to Image

Nano Banana 2.1

Krea 3 Image Edit

Krea 3 Image Edit (Image-to-Image)

What is Krea 3 Image Edit?

Key Features

Best Use Cases

Prompt Tips and Output Quality

How to Use Krea 3 Image Edit

Writing the prompt

Choosing settings

Recommended Settings

Examples to try

Reference-guided composites

How to restage a candle product photo in a lifestyle scene with Krea 3 Image Edit

How to recolor one object in a breakfast photo with Krea 3 Image Edit

How to virtually furnish an empty room photo from product shots with Krea 3 Image Edit

How to turn a product sketch into a studio product shot with Krea 3 Image Edit

Does it keep labels and logos intact?

How many reference images can I use?

What is regional edit?

What resolution does it output?

Related Models

Krea 3 Text to Image

Nano Banana 2.1

Nano Banana Pro

Segmind Faceswap v5