Krea 3 Text to Image: native 2K generation with a high thinking mode
Krea 3 Text to Image generates native 1K and 2K images from a text prompt in seconds, with a high thinking mode that plans complex, multi-element prompts before rendering.
Krea 3 Text to Image is the text-to-image endpoint of Krea 3, Krea's newest foundation image model. Give it a written prompt and it returns a finished image at native 1K or 2K in seconds. A high thinking mode plans complex, multi-element prompts before rendering, and the model holds a named medium instead of a generic look. Its sibling, Krea 3 Image Edit, starts from your own images; this endpoint generates from text alone.
ModelKrea 3 Text to ImageText-to-image at native 2K with a high thinking mode.What it does well
Plans a crowded prompt before it renders
Turn on high thinking mode and Krea 3 reasons about the whole prompt before it draws. In testing, a roller-derby jam with several interacting skaters, frozen motion, and three legible roster numbers landed in one clean native-2K frame. The bodies and hands stay correct where a fast single-subject pass tends to smear them. Keep high for any prompt that stacks figures, motion, and in-image text.
Roller derby jam on a banked wooden track: a jammer in a star-marked helmet cover bursts through a wall of three blockers, one blocker sliding on a knee pad. Roster numbers printed large on their arms and mesh tank jerseys read '07', '22', and '14'. Motion blur streaks off the spinning wheels while the skaters' bodies stay sharp; sweat, scuffed floor tape, taped-up pads. Direct on-camera flash, dark packed crowd behind, 35mm documentary sports photograph.
Holds a named medium
Name a medium and its materials, and Krea 3 keeps the output in that medium. A prompt for a ukiyo-e woodblock print came back with carved black outlines, flat indigo and vermilion ink, and visible paper grain — a real print, not a glossy photo of one. This art direction carries over from Krea 2, so paintings, vector art, and ink stay in their own medium. Stay at 2k so the fine line work holds.
Ukiyo-e woodblock print of a kabuki actor frozen in a mie pose, one arm thrust out, the sleeves of a crane-patterned kimono flaring; carved black outline, flat layers of indigo and vermilion ink, visible paper grain and a slight registration offset, a paper lantern at each edge.
Places every material in a dense brief
List one material per clause and Krea 3 places each where you put it. A mid-century living-room brief — burnt-orange wool sofa, teak credenza with a spinning record, travertine table, olive shag rug, brass arc lamp, paper lantern — arrived with every piece in its spot. At 2k, the small textures and the terrazzo floor stay sharp. Keep high so none of the named elements drop.
Mid-century modern living room photographed straight on: a burnt-orange wool sofa, a teak credenza with a record spinning on its turntable, a round travertine coffee table, an olive shag rug, a brass arc lamp curving overhead, and a white paper lantern lit in the corner; north-window daylight falling across a terrazzo floor.
Examples to try
Here is one more prompt that ran end to end. Click it to load the prompt and its exact settings into the form.
Studio fashion portrait of a woman turning to face the camera in a saffron silk sari with gold zari embroidery, a single softbox lighting the pleats, a seamless charcoal-grey backdrop, her hands resting at her waist.
Even at 1K with fast thinking, the portrait came back with an intact face and hands and a legible gold zari border. Draft single subjects here, then re-run a keeper at 2k.
Recommended settings
| Setting | Value | When |
|---|---|---|
resolution | 2k | Final deliverables, small text, or fine texture |
resolution | 1k | Fast, cheap iteration while you explore |
thinking_mode | high | Crowded, multi-element prompts |
thinking_mode | fast | Simple, single-subject prompts |
aspect_ratio | auto | Let the model frame; set a fixed ratio for a known canvas |
seed | empty | Leave empty for variety; a reused seed is not strictly reproducible |
Good to know
- Krea recommends Krea 3 for fast, instruction-following generation and editing, and Krea 2 for aesthetic exploration with style references, moodboards and LoRAs.
- A reused seed does not guarantee an identical image, because the prompt-expansion step varies between runs.
- LoRAs and style references are not available on this API yet.
- This endpoint generates from text only; to start from your own image, use the Krea 3 Image Edit sibling.
Use it from the API
Send a POST with your prompt and settings, and the response returns the image directly. The block below is the roller-derby example above.
# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind
result = segmind.run(
"krea-3-text-to-image",
prompt="Roller derby jam on a banked wooden track: a jammer in a star-marked helmet cover bursts through a wall of three blockers, one blocker sliding on a knee pad. Roster numbers printed large on their arms and mesh tank jerseys read '07', '22', and '14'. Motion blur streaks off the spinning wheels while the skaters' bodies stay sharp; sweat, scuffed floor tape, taped-up pads. Direct on-camera flash, dark packed crowd behind, 35mm documentary sports photograph.",
resolution="2k",
aspect_ratio="3:2",
thinking_mode="high",
)
print(result.get("output"))# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/krea-3-text-to-image" \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Roller derby jam on a banked wooden track: a jammer in a star-marked helmet cover bursts through a wall of three blockers, one blocker sliding on a knee pad. Roster numbers printed large on their arms and mesh tank jerseys read '\''07'\'', '\''22'\'', and '\''14'\''. Motion blur streaks off the spinning wheels while the skaters'\'' bodies stay sharp; sweat, scuffed floor tape, taped-up pads. Direct on-camera flash, dark packed crowd behind, 35mm documentary sports photograph.",
"resolution": "2k",
"aspect_ratio": "3:2",
"thinking_mode": "high"
}' | jq -r .request_id)
# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$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"Resources
- Try Krea 3 Text to Image on Segmind
- API reference
- Krea 3 API overview — Krea's developer reference: endpoint, parameters, prompting, and pricing.
- Krea 3 (model guide) — Krea's own guide to Krea 3: native 2K output, what to use it for, and Krea 3 vs Krea 2.
- Introducing Krea 3 — Krea's launch post with positioning, worked examples, and a recommended prompt pattern.
- Krea API: Text to Image examples — text-to-image usage and code samples across Krea models.
- #text-to-image
- #krea
- #model-launch
- #native-2k
- #thinking-mode
