Krea 2 Medium is a from-scratch Krea 2 image model designed for expressive illustration. Krea describes it as smaller than Large and more extensively post-trained, with a more stable character that works especially well for illustration, anime and painting. Medium and Large share the documented control system; the edition difference concerns their model design and visual tendencies rather than a different kind of output.
Use Medium to develop editorial illustrations, story scenes, painted concepts and coordinated visual directions. Start with a brief that identifies the subject, action, setting, composition, palette and medium. For example: “A cyclist crossing a quiet riverside bridge at dawn, wide editorial illustration, layered gouache shapes, soft blue shadows and one coral accent, open space above the river for a headline.” Treat the empty space as a composition request, then place final typography in the finished layout.
The generation interface requires a prompt, aspect ratio and resolution. Text supplies the scene and creative direction. An optional source image enables image-to-image generation, which starts from that image instead of pure noise. The denoising strength ranges from 0 to 1: 0 keeps the input image, 1 fully replaces it, and the documented default is 0.99. Use a source image when the starting composition is important, then describe the new treatment and select how much transformation to allow.
Image style references influence the visual treatment separately from the source image. The model-specific schema limits source-image and style-reference URI strings to 1,024 characters. Medium’s model-specific schema accepts up to ten references, with individual influence from 0 to 1 and a default of 0.5. Prepare a small, purposeful reference set: one image might establish brushwork and another the palette or lighting. State the intended relationship in the brief instead of relying on a mixed collection to imply a coherent direction.
A moodboard groups related references around a common style, palette, texture or composition. The documented workflow creates the moodboard first and then uses its existing identifier in a generation request. Medium accepts one moodboard, with influence from 0 to 1 and a default of 0.23 in its request schema. Trained styles, typically LoRAs, are a separate input selected by identifier and strength from -2 to 2. Choose the appropriate input for the job: a source image starts the scene, a reference suggests appearance, and a moodboard or trained style supplies a reusable direction.
Creativity controls expansion of the prompt. Raw disables expansion; low stays closer to the literal brief, while medium and high allow more interpretation. Medium’s model-specific request schema uses low as the default. For a tightly specified illustration, write the complete direction and start with restrained expansion. For early exploration, a looser setting can develop a simpler concept into a richer scene.
Three independent sliders shape the image: Intensity, Complexity and Movement. Each takes an integer from -100 to 100, with 0 neutral. Intensity shifts the treatment between muted and bold. Complexity shifts between clean, minimal compositions and dense detail. Movement shifts visual energy, gesture and implied motion. Movement remains a still-image control; it does not generate animation or a video clip. The sliders are independent of creativity’s prompt expansion and can be combined with reference inputs.
A practical illustration workflow establishes the scene first, then adjusts palette and treatment, then tunes density or energy. Keep the prompt, reference set and optional seed recorded while changing one control at a time. For the riverside illustration, lower complexity can simplify the background, while a higher movement value can suggest a livelier riding pose. These are creative directions to inspect in the completed result, not exact layout or pose locks.
Medium’s documented generation schema offers the 1K resolution scale and nine aspect ratios: 1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 3:4, 2:3 and 9:16. Select the intended frame before generating so the brief can describe its composition accordingly. A square illustration, a wide story scene and a vertical cover need different placement of the main subject and surrounding detail.
The documented generation flow returns a job, which may initially be pending. Follow its status to completion and then retrieve the resulting image locations. Preserve the selected image together with the brief, input images and control settings. Review the silhouette, hands or small objects, perspective, palette and any visible lettering before incorporating it into a publication or campaign.
Medium’s illustration emphasis is a useful edition choice, not a promise of identical results across prompts. A visual reference can guide style without perfectly copying its subject or layout, and image-to-image strength changes how much of the starting image survives. Use explicit instructions for required elements, inspect each completed output against the brief, and make final composition, typography and production adjustments in an editor.
", "aboutText": "### Expressive illustration and stable creative development\n\nKrea 2 Medium is a from-scratch Krea 2 image model designed for expressive illustration. Krea describes it as smaller than Large and more extensively post-trained, with a more stable character that works especially well for illustration, anime and painting. Medium and Large share the documented control system; the edition difference concerns their model design and visual tendencies rather than a different kind of output.\n\nUse Medium to develop editorial illustrations, story scenes, painted concepts and coordinated visual directions. Start with a brief that identifies the subject, action, setting, composition, palette and medium. For example: “A cyclist crossing a quiet riverside bridge at dawn, wide editorial illustration, layered gouache shapes, soft blue shadows and one coral accent, open space above the river for a headline.” Treat the empty space as a composition request, then place final typography in the finished layout.\n\n### Prepare text and image inputs\n\nThe generation interface requires a prompt, aspect ratio and resolution. Text supplies the scene and creative direction. An optional source image enables image-to-image generation, which starts from that image instead of pure noise. The denoising strength ranges from 0 to 1: 0 keeps the input image, 1 fully replaces it, and the documented default is 0.99. Use a source image when the starting composition is important, then describe the new treatment and select how much transformation to allow.\n\nImage style references influence the visual treatment separately from the source image. The model-specific schema limits source-image and style-reference URI strings to 1,024 characters. Medium’s model-specific schema accepts up to ten references, with individual influence from 0 to 1 and a default of 0.5. Prepare a small, purposeful reference set: one image might establish brushwork and another the palette or lighting. State the intended relationship in the brief instead of relying on a mixed collection to imply a coherent direction.\n\nA moodboard groups related references around a common style, palette, texture or composition. The documented workflow creates the moodboard first and then uses its existing identifier in a generation request. Medium accepts one moodboard, with influence from 0 to 1 and a default of 0.23 in its request schema. Trained styles, typically LoRAs, are a separate input selected by identifier and strength from -2 to 2. Choose the appropriate input for the job: a source image starts the scene, a reference suggests appearance, and a moodboard or trained style supplies a reusable direction.\n\n### Control interpretation and the visual result\n\nCreativity controls expansion of the prompt. Raw disables expansion; low stays closer to the literal brief, while medium and high allow more interpretation. Medium’s model-specific request schema uses low as the default. For a tightly specified illustration, write the complete direction and start with restrained expansion. For early exploration, a looser setting can develop a simpler concept into a richer scene.\n\nThree independent sliders shape the image: Intensity, Complexity and Movement. Each takes an integer from -100 to 100, with 0 neutral. Intensity shifts the treatment between muted and bold. Complexity shifts between clean, minimal compositions and dense detail. Movement shifts visual energy, gesture and implied motion. Movement remains a still-image control; it does not generate animation or a video clip. The sliders are independent of creativity’s prompt expansion and can be combined with reference inputs.\n\nA practical illustration workflow establishes the scene first, then adjusts palette and treatment, then tunes density or energy. Keep the prompt, reference set and optional seed recorded while changing one control at a time. For the riverside illustration, lower complexity can simplify the background, while a higher movement value can suggest a livelier riding pose. These are creative directions to inspect in the completed result, not exact layout or pose locks.\n\n### Format, generation and delivery\n\nMedium’s documented generation schema offers the 1K resolution scale and nine aspect ratios: 1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 3:4, 2:3 and 9:16. Select the intended frame before generating so the brief can describe its composition accordingly. A square illustration, a wide story scene and a vertical cover need different placement of the main subject and surrounding detail.\n\nThe documented generation flow returns a job, which may initially be pending. Follow its status to completion and then retrieve the resulting image locations. Preserve the selected image together with the brief, input images and control settings. Review the silhouette, hands or small objects, perspective, palette and any visible lettering before incorporating it into a publication or campaign.\n\nMedium’s illustration emphasis is a useful edition choice, not a promise of identical results across prompts. A visual reference can guide style without perfectly copying its subject or layout, and image-to-image strength changes how much of the starting image survives. Use explicit instructions for required elements, inspect each completed output against the brief, and make final composition, typography and production adjustments in an editor.", "editorialOverlay": { "copy_fields": null, "copy_sha256": null, "copy_source": null, "html_sha256": "f96af3c468003287d06d9e2b010c429ea36305e554d4dbaa74ca286442e573de", "module_sha256": "ee9dc1c50b1e6a8fb7fa062030fcb4646b9aaaeb08be513b723a5a92a7290f06", "raw_model_unchanged": true, "schema": "ailoot-native-model-editorial-v1" }, "locale": "en", "methods": [ { "html": "krea/krea-2-mediumKrea 2 Medium generates expressive illustrations from text and optional visual inputs. Its more extensive post-training favors stable results across illustration, anime and painting, with controls for style, prompt interpretation, composition density and visual energy.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Krea 2 Medium is a from-scratch Krea 2 image model designed for expressive illustration. Krea describes it as smaller than Large and more extensively post-trained, with a more stable character that works especially well for illustration, anime and painting. Medium and Large share the documented control system; the edition difference concerns their model design and visual tendencies rather than a different kind of output.
Use Medium to develop editorial illustrations, story scenes, painted concepts and coordinated visual directions. Start with a brief that identifies the subject, action, setting, composition, palette and medium. For example: “A cyclist crossing a quiet riverside bridge at dawn, wide editorial illustration, layered gouache shapes, soft blue shadows and one coral accent, open space above the river for a headline.” Treat the empty space as a composition request, then place final typography in the finished layout.
The generation interface requires a prompt, aspect ratio and resolution. Text supplies the scene and creative direction. An optional source image enables image-to-image generation, which starts from that image instead of pure noise. The denoising strength ranges from 0 to 1: 0 keeps the input image, 1 fully replaces it, and the documented default is 0.99. Use a source image when the starting composition is important, then describe the new treatment and select how much transformation to allow.
Image style references influence the visual treatment separately from the source image. The model-specific schema limits source-image and style-reference URI strings to 1,024 characters. Medium’s model-specific schema accepts up to ten references, with individual influence from 0 to 1 and a default of 0.5. Prepare a small, purposeful reference set: one image might establish brushwork and another the palette or lighting. State the intended relationship in the brief instead of relying on a mixed collection to imply a coherent direction.
A moodboard groups related references around a common style, palette, texture or composition. The documented workflow creates the moodboard first and then uses its existing identifier in a generation request. Medium accepts one moodboard, with influence from 0 to 1 and a default of 0.23 in its request schema. Trained styles, typically LoRAs, are a separate input selected by identifier and strength from -2 to 2. Choose the appropriate input for the job: a source image starts the scene, a reference suggests appearance, and a moodboard or trained style supplies a reusable direction.
Creativity controls expansion of the prompt. Raw disables expansion; low stays closer to the literal brief, while medium and high allow more interpretation. Medium’s model-specific request schema uses low as the default. For a tightly specified illustration, write the complete direction and start with restrained expansion. For early exploration, a looser setting can develop a simpler concept into a richer scene.
Three independent sliders shape the image: Intensity, Complexity and Movement. Each takes an integer from -100 to 100, with 0 neutral. Intensity shifts the treatment between muted and bold. Complexity shifts between clean, minimal compositions and dense detail. Movement shifts visual energy, gesture and implied motion. Movement remains a still-image control; it does not generate animation or a video clip. The sliders are independent of creativity’s prompt expansion and can be combined with reference inputs.
A practical illustration workflow establishes the scene first, then adjusts palette and treatment, then tunes density or energy. Keep the prompt, reference set and optional seed recorded while changing one control at a time. For the riverside illustration, lower complexity can simplify the background, while a higher movement value can suggest a livelier riding pose. These are creative directions to inspect in the completed result, not exact layout or pose locks.
Medium’s documented generation schema offers the 1K resolution scale and nine aspect ratios: 1:1, 4:3, 3:2, 16:9, 2.35:1, 4:5, 3:4, 2:3 and 9:16. Select the intended frame before generating so the brief can describe its composition accordingly. A square illustration, a wide story scene and a vertical cover need different placement of the main subject and surrounding detail.
The documented generation flow returns a job, which may initially be pending. Follow its status to completion and then retrieve the resulting image locations. Preserve the selected image together with the brief, input images and control settings. Review the silhouette, hands or small objects, perspective, palette and any visible lettering before incorporating it into a publication or campaign.
Medium’s illustration emphasis is a useful edition choice, not a promise of identical results across prompts. A visual reference can guide style without perfectly copying its subject or layout, and image-to-image strength changes how much of the starting image survives. Use explicit instructions for required elements, inspect each completed output against the brief, and make final composition, typography and production adjustments in an editor.
POST /v1/images/generationstext, imageimage| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits krea/krea-2-medium | Default— |
Fieldprompt | Typestring | RequirementRequired | Values & limits Minimum length: 1 | Default— |
Fieldresolution | Typestring | RequirementOptional | Values & limits 1K | Default— |
Fieldaspect_ratio | Typestring | RequirementOptional | Values & limits 1:1, 4:3, 3:2, 16:9, 4:5, 2:3, 9:16 | Default— |
Fieldinput_references | Typearray | RequirementOptional | Values & limits Minimum items: 0 Maximum items: 1 | Default— |
Fieldprovider | Typeobject | RequirementOptional | Values & limits | Default— |
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, prompt |
Parametermaterial_types | Contracttext, image |
Lifecycle: synchronous_result_then_owned_library
GET /v1/media/tasks/{task_id} · actual returned task_id or media library record
GET /v1/media/assets/{asset_id}/content
GET /v1/media/assets
GET /v1/media/assets/{asset_id}
The displayed rates estimate the cost before generation. The final charge is calculated from the cost recorded for the original generation. Measured output duration and forecast quantities are not the final billable quantity.
Tariff references do not add request options. Supported parameters are listed for the selected method.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Forecast quantity | Charge |
|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Forecast quantity— | Charge Input · Input tokens · Estimate for comparison only — / 1K tokens Output · Output tokens · Estimate for comparison only — / 1K tokens Input · Cache read · Estimate for comparison only — / 1K tokens Input · Cache write · Estimate for comparison only — / 1K tokens Input · Cache write, 1 hour · Estimate for comparison only — / 1K tokens |
| ConfigurationEstimated generation cost | Rate schedule— | Forecast quantityone generation estimate | Charge Output · Estimated generation cost · Estimate for comparison only $0.06 estimate before generation This estimate is not an additional charge or a fixed rate for the completed generation. |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\nexport AILOOT_API_KEY\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np={'model': 'krea/krea-2-medium', 'prompt': 'a red panda astronaut floating in space, studio lighting'}\n\nPath("request.json").write_text(json.dumps(p,ensure_ascii=False))\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/images/generations \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\npython3 - <<'PY_RESULT'\nimport base64, json, os, subprocess\nfrom pathlib import Path\nfrom urllib.parse import urlsplit, urljoin\n\ndef save_images(items):\n if not items: raise SystemExit("No completed images; retain the full response")\n for index, item in enumerate(items):\n encoded=item.get("b64_json")\n uri=item.get("url", "")\n if not encoded and uri.startswith("data:image/"):\n header, separator, encoded=uri.partition(",")\n if not separator or ";base64" not in header: raise SystemExit("Unsupported image data URI")\n if encoded:\n raw=base64.b64decode(encoded,validate=True)\n elif uri:\n uri=urljoin("https://api.ailoot.ai",uri)\n parsed=urlsplit(uri)\n if parsed.scheme!="https" or not parsed.netloc or parsed.username or parsed.password:\n raise SystemExit("Invalid returned image URL")\n owned=parsed.netloc=="api.ailoot.ai" and parsed.path.startswith("/v1/media/assets/") and parsed.path.endswith("/content")\n args=["curl","--fail-with-body","--silent","--show-error",uri]\n if owned: args += ["-H","Authorization: Bearer "+os.environ["AILOOT_API_KEY"]]\n temporary=Path(f"image-{index}.bin")\n subprocess.run(args+["-o",str(temporary)],check=True)\n raw=temporary.read_bytes()\n temporary.unlink()\n else: raise SystemExit("Image has neither bytes nor URL")\n if not raw: raise SystemExit("Empty image result")\n ext="png" if raw.startswith(b"\\x89PNG") else "jpg" if raw.startswith(b"\\xff\\xd8") else "webp" if raw[:4]==b"RIFF" and raw[8:12]==b"WEBP" else "svg" if raw.lstrip().startswith((b"<svg",b"<?xml")) else "bin"\n Path(f"image-{index}.{ext}").write_bytes(raw)\n print("Every returned image saved; complete response metadata retained")\nsave_images(json.loads(Path("result.json").read_text())["data"])\nPY_RESULT\nDownload each returned image or save its Base64 contents
GET /v1/media/tasks/{task_id}GET /v1/media/assets/{asset_id}/contentGET /v1/media/assetsGET /v1/media/assets/{asset_id}set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/media/assets \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -o owned-assets.json\n# Preserve every returned asset and follow its exact signed content URL. Do not construct a signed token.\nSave every completed image; retrieve each returned URL or decode its Base64 contents.
Retain the complete response and headers, every actual output, original returned identifiers and all usage fields. Recover the original operation without submitting creation again.
Keep the returned X-Client-Request-ID for reconciliation. Read only actual returned task_id/asset_id values or the owned media library; never build a task URL from the operation ID and never resubmit creation.
| Field | Type |
|---|---|
Fieldcreated | Type |
Fielddata[].url | Type |
Fielddata[].b64_json | Type |
Fieldusage | Type |
Ailoot Ultra krea/krea-2-medium
/v1/images