by Kuaishou

Native 2K, 4K or 8K stills with readable lettering, plus edits from up to ten references

kling/image-4.0/text-to-image
  • Up to 8K
  • Up to 10 reference images
  • 8 aspect ratios
  • Up to 9 per run
  • Create Image
  • Edit Image

About Kling Image 4.0

Native 2K, 4K or 8K stills with readable lettering, plus edits from up to ten references.

Call it with the model name kling/image-4.0/text-to-image on POST /jobs/createTask, from $0.063.

Introduction

Turn a written brief into a still image

Use this endpoint when the scene can be described without an uploaded source picture. Kling Image 4.0 takes a required prompt and offers 2k, 4k, and 8k output tiers, with one to nine independent images per request. Our runs include a tea festival poster, a harbour at dawn, and an ink-and-watercolour fern.

A batch gives independent alternatives from the same instruction. Choose an output count that gives you a manageable set to review, and use the related Set family when the deliverable needs a coordinated collection.

Request controls

The prompt is the required field. The other controls set the output count, resolution tier, and shape; leaving them out uses the defaults below.

ParameterAccepted inputDefaultPractical role
promptRequired string, 1 to 21,000 charactersNoneDescribe the subject, setting, lighting, and exact words to draw.
num_outputsInteger from 1 to 91Request independent images from the same brief.
aspect_ratioauto, 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9autoChoose a fixed frame or let the model interpret the brief.

Choose the resolution tier for the intended delivery and the aspect ratio for the placement. Check the returned file dimensions when preparing the image for your layout; the examples below show the sizes returned by our requests.

Give the model a composition to draw

Start with a visible subject and its setting, then specify the light and visual treatment. The harbour brief names wooden fishing boats, dawn mist, and pale gold light. The fern brief names a plant, off-white paper, fine ink lines, and muted green watercolour. Each describes what should be visible rather than relying on general praise for image quality.

Make lettering a separate instruction

For a poster, quote the intended title and locate it in the frame. Proofread the full returned image, including smaller lines you did not request.

Kling says a picture can contain up to 2,000 characters of generated text. Keep this separate from the 21,000-character prompt limit: one concerns the lettering drawn in the picture, while the other covers your full description. Group titles and supporting lines by their intended position so the layout instructions stay readable.

Align the scene with the frame

Choose a fixed ratio when the destination is known: a tall poster and a wide editorial scene have different framing needs. Use auto when you want the prompt to guide the shape. If you change the resolution after reviewing a draft, inspect the new image again rather than assuming that the earlier composition and lettering carried over unchanged.

What our November 18 runs returned

Our three text-to-image runs used different subjects, resolution tiers, and frame shapes. Review the examples for composition, surface treatment, and lettering as you plan your own image.

Tea poster at 2k

The SLOW STEEP request used a 3:4 frame and returned one 1760 x 2336 PNG. The title was spelled correctly above an indigo engraved teapot on cream paper. The model also invented supporting lines. Check the supporting text as well as the headline before using the poster.

Harbour at 4k

The dawn harbour request used 16:9 and returned one 5440 x 3072 PNG showing wooden boats, mist, and pale dawn light. The wide frame gives the boats, mist, and surrounding harbour space within the composition.

Fern at 8k

The square botanical request returned one 8192 x 8192 PNG of an ink-and-watercolour fern on off-white paper. Kling says its 8k tier uses native generation. Inspect the example for its fine line work, colour treatment, and square framing.

Scope and related families

This endpoint accepts text rather than a source image.

Kling says series generation coordinates related images and long-image generation makes product-detail or webpage visuals. Choose Kling Image 4.0 Set (kling/image-4.0-set) for a coordinated collection, or Kling Image 4.0 Long (kling/image-4.0-long) for that page-shaped deliverable. The independent still-image controls here serve a different output structure. When comparing candidates, use the same intended composition, exact copy, and delivery requirements, then review the images at the size you plan to use.

Request parameters

ParameterRequiredDescription
promptYesDescribe what you want to make.
resolutionNoChoose an available output size. 2k · 4k · 8k · Default: 2k
num_outputsNoHow many results to make in one go. ≥ 1 · ≤ 9 · Default: 1
aspect_ratioNoThe shape of the frame: wide, square or tall. auto · 16:9 · 9:16 · 1:1 · 4:3 · 3:4 · 3:2 · 2:3 · 21:9 · Default: auto

Pricing

Prices are in USD and charged to your PopNow balance. When a task is accepted, we reserve the maximum quoted amount from your API balance. On completion, we charge for actual usage, never more than the hold, and release the remainder. If the task fails, the hold is released.

OptionUSD per image
resolution 2K$0.063
resolution 4K$0.094
resolution 8K$0.157

Send the same request body to POST /jobs/quote to see the estimated and maximum USD charge before running. A quote does not run the task or charge your balance.

API quick start

Call the Kling Image 4.0 API in three steps

  1. Create an API key. Sign in to PopNow and create a key in the API console. Keys start with sk-popnow-. API calls are billed in USD from your PopNow balance.
  2. Send the request. POST /jobs/createTask with model kling/image-4.0/text-to-image and an input object that follows the parameter table. The response returns a task ID straight away.
  3. Get the result. Poll GET /jobs/recordInfo with the task ID, or pass a callBackUrl and we post the result to it. Download the output links when the task succeeds.
# 1. Create the task. Keep the Idempotency-Key if you retry after a network error.
export IDEMPOTENCY_KEY="$(uuidgen)"
curl -X POST https://api.popnow.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $POPNOW_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $IDEMPOTENCY_KEY" \
  -d '{
  "model": "kling/image-4.0/text-to-image",
  "input": {
    "prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
    "num_outputs": 1,
    "resolution": "2k",
    "aspect_ratio": "auto"
  }
}'

# 2. Copy data.taskId from the response, then poll until state is succeeded, failed, canceled or expired.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.popnow.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $POPNOW_API_KEY"

Replace the prompt and any links with your own. The API tab has the response, polling and callback reference.

Kling Image 4.0 API questions

How much does the Kling Image 4.0 API cost?

Kling Image 4.0 starts at $0.063 per image with Create Image, its lowest-priced task. Prices by option: 2K $0.063 per image, 4K $0.094 per image, 8K $0.157 per image. API usage is charged in USD to your PopNow balance. When a task is accepted, we reserve the maximum quoted amount from your API balance. On completion, we charge for actual usage, never more than the hold, and release the remainder. If the task fails, the hold is released.

What is the model name for Kling Image 4.0 in the API?

Put kling/image-4.0/text-to-image for Create Image; kling/image-4.0/edit for Edit Image in the model field of POST /jobs/createTask.

Does Kling Image 4.0 really output 8K?

Yes, 8k is an available output tier. Our square fern run returned an 8192 x 8192 PNG of fine ink lines and muted green watercolour on off-white paper. Kling says this tier uses native generation.

Are the nine images in one request related?

No, they are independent takes on the same prompt. Kling says coordinated image collections use its series workflow; choose the related Kling Image 4.0 Set family for that purpose.

How do I use the Kling Image 4.0 API for a still image?

Choose text to image and provide a prompt. Set the resolution, aspect ratio, and output count when the defaults do not fit your layout. Image editing also requires one to ten reference images, with numbered citations in the instruction.

Can it add text that I did not request?

Yes, both our tea poster and rose edit added unrequested words. The poster spelled SLOW STEEP correctly but invented supporting lines; the rose edit added ROSE and NATURAL ESSENCE. Quote the intended copy, add "no other words," and proofread the returned image.

How long can the prompt be?

Both endpoints accept 1 to 21,000 characters. Use that space to describe the subject, composition, lighting, exact copy, and any reference roles. Keep the requested changes together and remove conflicting directions before submitting.

What are the limits of the Kling Image 4.0 API?

Resolutions: 2K · 4K · 8K; Aspect ratios: auto · 16:9 · 9:16 · 1:1 · 4:3 · 3:4 +3; Outputs per call: 1–9; Prompt length (characters): 21,000. Requests outside these limits are rejected before anything is charged.

Can I try Kling Image 4.0 without writing code?

Run Kling Image 4.0 in the Playground with your API USD balance, without an API key. Each run appears in your developer history with its matching code.

All Kling models

More about Kling Image 4.0