kling-3-prompting（kling 3 提示）

Overview

Kling 3.0 is a unified multimodal video model. It understands cinematic direction, not keyword lists. Write prompts like a director — describe what the audience sees, hears, and feels over time.

Core shift: Description → Direction. Think "direct a scene" not "describe an image."

Interactive Builder Workflow

When invoked, guide the user through these steps using AskUserQuestion:

digraph builder {
  "1. Generation mode?" [shape=diamond];
  "Text-to-Video" [shape=box];
  "Image-to-Video" [shape=box];
  "Multi-Shot Sequence" [shape=box];
  "Keyframe Transition" [shape=box];
  "2. Gather scene details" [shape=box];
  "3. Assemble prompt" [shape=box];
  "4. Present & refine" [shape=box];

  "1. Generation mode?" -> "Text-to-Video";
  "1. Generation mode?" -> "Image-to-Video";
  "1. Generation mode?" -> "Multi-Shot Sequence";
  "1. Generation mode?" -> "Keyframe Transition";
  "Text-to-Video" -> "2. Gather scene details";
  "Image-to-Video" -> "2. Gather scene details";
  "Multi-Shot Sequence" -> "2. Gather scene details";
  "Keyframe Transition" -> "2. Gather scene details";
  "2. Gather scene details" -> "3. Assemble prompt";
  "3. Assemble prompt" -> "4. Present & refine";
}

Step 1: Determine Generation Mode

Ask the user which mode:

• Text-to-Video — prompt from scratch
• Image-to-Video — animate a reference image
• Multi-Shot Sequence — 2-6 shot storyboard (up to 15s)
• Keyframe Transition — start frame → end frame with interpolated motion

Step 2: Gather Scene Details

Ask about each element (adapt questions to mode):

Image-to-Video: Focus on how the scene *evolves from* the image — movement, camera motion, environmental change. The model preserves identity/layout from the source.

Keyframes: Ask for start and end frame descriptions. Frames should match in color, style, and lighting. Prompt sparingly — Kling infers motion well.

Multi-Shot: Define each shot separately with its own framing, subject, action, and duration. Label shots explicitly.

Step 3: Assemble the Prompt

Use the Master Formula:

[Scene/Environment] + [Subject & Appearance] + [Action Timeline] + [Camera Movement] + [Audio & Atmosphere] + [Technical Specs]

Writing rules:

• Use cinematic motion verbs: dolly push, whip-pan, crash zoom, rack focus, tracking shot — NOT "moves" or "goes"
• Name real light sources: neon signs, candlelight, golden hour, LED panels — NOT "dramatic lighting"
• Include texture for credibility: grain, lens flares, condensation, fabric sheen, smoke, sweat
• Describe temporal flow: beginning → middle → end
• Keep to 1-3 rich sentences per shot (specificity > length)
• For dialogue: use character labels, assign voice tone/emotion, use transitional words ("Immediately," "Pause")

Step 4: Present & Refine

Present the assembled prompt. Ask if they want to:

• Adjust any element
• Add a negative prompt
• Generate variations (different duration, different camera, different mood)

Quick Reference

Camera Movements

Lens & Film Stock

Lighting

Use specific sources, not adjectives:

• "Golden hour sun cutting through dusty warehouse windows"
• "Flickering neon casting magenta/cyan across wet pavement"
• "Single bare bulb swinging, casting moving shadows"
• "Cool blue LED panels reflecting off glass surfaces"
• "Candlelight warming skin tones, deep shadows beyond"

Color & Grade

• "Desaturated teal grade, crushed blacks"
• "Amber nightclub strobe cutting through smoke"
• "Cool blue haze filling the corridor"
• "Magenta neon reflecting off wet asphalt"
• "Overexposed highlights, blown-out whites"

Multi-Character Dialogue

Multi-Shot Structure

Shot 1 (0-5s): [Wide establishing shot description]
Shot 2 (5-10s): [Medium/close-up with action progression]
Shot 3 (10-15s): [Resolution/reaction with camera payoff]

Atmosphere: [Overall mood, color grade]
Audio: [Sound design, music, dialogue]

Label every shot. Assign durations. Describe framing + subject + motion per shot.

Start & End Frame Tips

• Frames should match in color palette, style, and lighting
• Identical start/end frames = seamless loop
• Prompt sparingly — Kling infers motion between frames well
• Simple camera directions: zoom in/out, pan left/right, tilt up/down
• 5s for dynamic transitions, 10s for complex transformations
• Start frame aspect ratio drives the whole clip

Negative Prompts

Use to prevent common AI defaults:

smiling, laughing, cartoonish, bright saturated colors, low resolution,
morphing, blurry text, disfigured hands, extra fingers, static pose,
frozen expression, stock photo aesthetic

Customize based on scene — remove items that conflict with your intent.

Weak → Strong

Common Mistakes