Kling 3.0 Image

Reference Image (Optional, 0-1)

Click or drag a file to this area to upload
Format: jpg/jpeg/png, Size: up to 10M
Min 300px, ratio 1:2.5~2.5:1

Prompt0/2000
Aspect Ratio9:16
Resolution1K
Image Count1
Paid User OnlyLearn More & Upgrade β€Ί
Kling 3.0 Image User Guide
Model Highlights
Visual Chain-of-Thought (vCoT)Kling 3.0 uses visual chain-of-thought reasoning to analyze scene structure, composition, and cinematic standards before rendering, significantly improving texture mapping and lighting physics while reducing common AI artifacts.
Native High-Resolution OutputSupports native 1K / 2K resolution output without upscaling, producing crisp and detailed images suitable for professional creative work and asset production.
Reference-Based GenerationUpload 1 reference image for image-to-image generation, maintaining consistency in composition, color palette, and style. For multi-image reference (up to 10), use the Kling Omni feature.
Batch GenerationGenerate 1, 2, or 4 images per request with unified style, atmosphere, and character elements across frames β€” ideal for sequential storytelling and series creation.
Basic Operations
1
Write Your Prompt
Describe your desired image in detail in the prompt box. Supports both Chinese and English, up to 2000 characters. Include subject, setting, lighting, mood, and style for best results.
2
Select Aspect Ratio
Available ratios: 1:1 (square), 16:9 (widescreen), 9:16 (portrait), 4:3, 3:4, 3:2, 2:3, 21:9 (ultrawide). Default is 9:16.
3
Choose Resolution
Select 1K standard or 2K high definition. 2K offers richer details, ideal for high-quality output. Default is 1K.
4
Add Reference Images (Optional)
Upload 1 reference image to enter image-to-image mode. The model will reference the composition, color palette, and style. Skip this for pure text-to-image generation. For multi-image reference, use Kling Omni.
5
Set Image Count
Choose how many images to generate per request: 1, 2, or 4. Cost = 20 credits Γ— number of images.
6
Generate
Click "Generate Image" to submit. Generation typically takes 30-120 seconds. View and download results in "My Result" when complete.
Prompt Tips
Structured DescriptionOrganize prompts as: Scene β†’ Characters β†’ Action β†’ Camera β†’ Style. Example: "On a rain-soaked Tokyo street (scene), a woman in a black trenchcoat (character), looking back while holding a clear umbrella (action), medium shot from the side (camera), cyberpunk style with neon reflections (style)".
Be SpecificAvoid vague descriptions like "cinematic lighting". Use specific details like "neon signs casting magenta and cyan reflections on wet asphalt". The more precise your physical descriptions, the better the results.
Use Negative PromptsSpecify unwanted elements in the negative prompt field, such as "blurry, low quality, deformed, watermark, text". Note: Negative prompts do not take effect in image-to-image mode.
Important Notes
Image-to-Image LimitationWhen reference images are uploaded, negative prompts are automatically disabled. Remove all reference images if you need to use negative prompts.
Reference Image RequirementsReference images should be at least 300Γ—300 pixels with an aspect ratio between 1:2.5 and 2.5:1. Supported formats: JPG / JPEG / PNG, max 10MB per file.
PricingFixed rate of 20 credits per image. Batch cost = 20 Γ— number of images. Credits are automatically refunded if generation fails.