Kling Image O3 Image To Image

From 0 credits/call
ae_kling_image_o3_image_to_image

Create images from a text prompt and up to 10 reference images, with optional character or object control, related image series, and resolutions up to 4K.

Image
From 0 credits/call

Pricing

From 0 credits/call
ConditionCharge
image resolution is 1K33.6 AE credits per item, based on confirmed billable images.
image resolution is 2K33.6 AE credits per item, based on confirmed billable images.
image resolution is 4K67.2 AE credits per item, based on confirmed billable images.
image resolution not supplied33.6 AE credits per item, based on confirmed billable images.

1K, 2K or omitted resolution: 33.6 AE credits per confirmed billable image; 4K: 67.2 AE credits per billable image. Confirmed zero usage is charged 0.

Billing uses confirmed billable image count. Missing, invalid or mismatched usage remains pending verification; it is not treated as zero.

Parameters

Optional parameters use the platform defaults when omitted.

ParameterTypeRequiredDescriptionDefault
request_bodyobjectYesNo parameter description.—
request_body.promptstringYesText prompt for image generation. Reference images using @Image1, @Image2, etc. (or @Image if only one image). Max 2500 characters.—
request_body.elementsarray | unknownNoOptional: Elements (characters/objects) for face control. Reference in prompt as @Element1, @Element2, etc.—
request_body.elements[].frontal_image_urlstring | unknownNoThe frontal image of the element (main view). Optional.—
request_body.elements[].reference_image_urlsarray | unknownNoAdditional reference images from different angles. 0-3 images supported. Optional.—
request_body.sync_modebooleanNoIf `True`, the media will be returned as a data URI.false
request_body.image_urlsarrayYesList of reference images. Reference images in prompt using @Image1, @Image2, etc. (1-indexed). Max 10 images.—
request_body.num_imagesintegerNoNumber of images to generate (1-9). Only used when result_type is 'single'.1
request_body.resolutionenum (1K | 2K | 4K)NoImage generation resolution. 1K: standard, 2K: high-res, 4K: ultra high-res.1K
request_body.result_typeenum (single | series)NoResult type. 'single' for one image, 'series' for a series of related images.single
request_body.aspect_ratioenum (16:9 | 9:16 | 1:1 | 4:3 | 3:4 | 3:2 | 2:3 | 21:9 | auto)NoAspect ratio of generated images. 'auto' intelligently determines based on input content.auto
request_body.output_formatenum (jpeg | png | webp)NoThe format of the generated image.png
request_body.series_amountinteger | unknownNoNumber of images in series (2-9). Only used when result_type is 'series'.—

Invocation example

Copy the request and explore its response example.

POST/agent-api/v1/tool/executeX-Api-Key

Submit first, then query with the returned task_id until the task completes.

Request coderequest.sh
curl --request POST \
  --url 'https://agentearth.ai/agent-api/v1/tool/execute?tool_name=ae_kling_image_o3_image_to_image' \
  --header "X-Api-Key: $AGENTEARTH_API_KEY" \
  --header 'Content-Type: application/json' \
  --data-raw '{
  "params": {
    "request_body": {
      "aspect_ratio": "3:4",
      "image_urls": [
        "https://agentearth.ai/fileserver/file/bcd3df79-f952-484e-a14a-b024b8cdf6a4.png",
        "https://agentearth.ai/fileserver/file/a8b1b6df-414a-40e8-bd5c-59fac38ab69b.jpg",
        "https://agentearth.ai/fileserver/file/7bb1877e-6682-429f-a50f-82bc1ad5cb92.jpg"
      ],
      "num_images": 1,
      "output_format": "png",
      "prompt": "NO TEXT/LETTERS/LOGO/SIGNATURE/WATERMARK anywhere; refs contain typography, copy none.\n\n@Image1 = POSE-ONLY lineart: copy ONLY pose, head tilt, hand gestures; ignore its monochrome sketch look, bald head, rough strokes. @Image2 @Image3 = STYLE-ONLY: borrow drawing method, never their characters, caps, outfits, props.\n\nSTYLE: clean hand-drawn black ink lineart, crisp strokes, variable weight (thicker outer contours/hems/shadow sides, thinner facial lines), tapered ends, a few broken dry-brush gaps. Flat two-tone cel color, VISIBLE hard-edged solid shadow blocks under chin, under sleeves, under hem, inside hair, along one leg; no gradient/airbrush/soft shading/3D/gloss/texture/denim wash. Pale ivory skin mostly unshaded, soft pink blush; big anime eyes with lashes, tiny star/triangle marks under eyes, pointed elf ears, calm bored asymmetric expression. Only a FEW contours have a hairline-thin subtle red offset line on one side, cyan on the other - never a full outline. One dominant saturated hue + one accent over white/pale grey/cream halftone paper; large empty space, 1-2 tiny original motifs, one thin ribbon; 3:4 vertical, flat even light.\n\nDENSITY: many clothing layers + many small accessories, EACH a large simple flat color block; information from the NUMBER of distinct items and color divisions, NOT fine lines. No fold lines/hatching/individual hair strands/lace/stitching.\n\nDESIGN 3: dark violet messy low bun with a plain ribbon and loose strands blown sideways, rose-pink eyes, teal star marks under eyes. POSE: half-body to upper thigh, RIGHT arm stretched fully straight up-right with wrist near frame edge, other arm bent back behind the waist, torso bent strongly sideways left so the shoulder line is steep and hips shift opposite way, head tilted left, chin slightly lifted, looking at viewer, legs cropped at upper thigh, hair and a long thin scarf streaming the same way. No extreme foreshortening. Rich layers: white strappy camisole, oversized off-shoulder tee with one plain chest pocket and contrast collar trim, dark cargo shorts with two large patch pockets, thin belt, thigh strap, thin-soled low-top shoes. Many chunky flat accessories: 4 star/arrow hair pins, bun ribbon, hoops + ear cuff, 2-layer necklace, choker, beaded bracelet on raised wrist, 2 rings, wristwatch, enamel pin, bag charm. Palette: soft lavender dominant, vivid coral orange accent, pale grey/cream ground. One thin diagonal ribbon, one tiny star.",
      "resolution": "1K",
      "result_type": "single",
      "sync_mode": false
    }
  }
}'
Response exampleresponse.json
JSON
{
  "description": "task accepted; poll the result via query_tool with this task_id",
  "query_tool": "ae_get_result",
  "status": "created",
  "task_id": "07280d59-d4a8-41a3-9401-3f7b49075ac8"
}

Guides using this tool

See how this tool fits into real workflows