Kling 3.0 Multi-Shot Prompts: Storyboard 6 Shots in 15s

Before Kling 3.0, making a short scene with cuts meant generating each shot on its own and hoping the character looked the same in all of them. Kling 3.0 changed that. It can generate several shots in one generation, with real camera cuts between them, while keeping the same characters and setting from shot to shot.

It is also easy to waste. Ask for six shots in five seconds and you get a stuttering mess. Leave the model to plan the cuts on its own and it may cut away from the moment you actually cared about. This guide explains how Kling 3.0’s multi-shot modes work, how to budget time across shots, how to write each shot, and gives you eight storyboard prompts ready to adapt.

If you are new to Kling in general, read our Kling AI prompt guide first for the basic scene, character, action and camera structure. This post assumes you know that and focuses on sequences.

What Multi-Shot Actually Is

Kling released VIDEO 3.0 in February 2026, alongside VIDEO 3.0 Omni. According to Kling’s own guides, the key numbers are:

• Up to six shots in one generation
• 3 to 15 seconds total, with the shot lengths adding up to the total
• 720p or 1080p output
• Native audio, including multi-character dialogue

Each shot must be at least one second long. A faster variant, Kling 3.0 Turbo, arrived in June 2026 with the same multi-shot storyboarding, built for quicker iteration.

There are two modes:

Multi-Shot (automatic). You write one prompt describing the scene. Kling plans the cuts itself, using standard film conventions to decide where the shots break. Good for quick drafts and when you do not care exactly where the cuts land.

Custom Multi-Shot. You define each shot yourself: how many, how long each one runs, the framing, the camera angle and what happens. Some platforms that host Kling call these Smart Storyboard and Custom Storyboard. Use this mode whenever timing matters, which for most content is always.

Budget Your Seconds First

The most common multi-shot mistake is cutting too often. More shots means less time per shot. Six shots in 15 seconds leaves about 2.5 seconds each, which is fine for a fast montage but not enough for a character to do anything meaningful.

A simple way to plan:

Total length	Shots	What fits
5s	2	Setup and payoff (a look, then a reaction)
8 to 10s	3	Establish, action, reaction
10 to 12s	4	A short scene with one line of dialogue
15s	4 to 6	A full mini-story or montage

Give the longest duration to the shot that carries the story. A reaction shot can be one or two seconds. The moment the glass hits the floor needs three.

Kling’s own guidance for durations is similar: a few seconds for a single action like a smile, around 10 seconds for simple dialogue, and the full 15 seconds for complex scenes with several cuts. When in doubt, start with two or three shots and add more once those work.

How to Write Each Shot

Kling’s example format is a plain numbered list with the duration after the shot number:

Shot 1, 4s: A wide tracking shot of a mountain range. Shot 2, 6s: A close-up of a climber reaching for a ledge. Shot 3, 5s: An aerial view of the summit.

Each shot line should answer three questions:

1. Framing: wide, medium, close-up, over-the-shoulder, insert
2. Camera: static, slow dolly push, handheld, tracking (one move per shot)
3. Action: one clear thing that happens

Then add a line before the shots that describes what stays the same: the characters, the setting, the lighting and the style. That shared line is what keeps the whole sequence consistent.

Shot Grammar That Works

Real editors cut in patterns that audiences understand without thinking. Kling’s multi-shot mode responds well to the same patterns:

• Wide, then closer. Show where we are, then show who matters. An establishing shot followed by a medium shot is the most natural opening there is.
• Action, then reaction. Something happens, then we see someone’s face respond. This is how comedy and horror both land.
• Insert shots. A one-second close-up of a hand, a phone screen or a door handle gives the viewer the key detail.
• Keep screen direction. If a character exits frame right in shot 2, they should enter from the left in shot 3. Say so in the prompt.
• Sound across the cut. A sound that starts in one shot and continues into the next makes the cut feel intentional.

Our AI video transitions guide goes deeper on match cuts and whip pans, which also work inside a multi-shot generation.

Dialogue and Characters

Kling 3.0 generates dialogue with lip-sync inside multi-shot sequences. Kling’s guide shows the speaker format as a character name with tone and language in parentheses, then the line in quotes:

Maya (nervous, English): "Is anyone else hearing that?"

Tag every speaker explicitly when more than one person is in the scene. Otherwise lines can come out of the wrong character.

For consistency across shots, and across multiple generations, use Elements. With VIDEO 3.0 Omni, you reference an element in the prompt with the @ symbol (for example @Maya), and you can build an element from a 3 to 8 second video clip. Kling says it extracts both the look and the original voice from that clip, so the same character sounds the same in every shot. One limit to know: Kling notes that native audio is not yet supported when you use video input with Omni.

For the wider picture on keeping characters stable across a series, see our consistent characters guide.

8 Copy-Ready Kling 3.0 Multi-Shot Prompts

Use these in Custom Multi-Shot mode. Change the characters and setting, keep the timing and shot structure.

1. Pet comedy (3 shots, 8s)

A small tidy kitchen in bright morning light. A chubby orange cat and a full glass of milk on the counter. Realistic, handheld phone video look.
Shot 1, 3s: Medium shot, static. The cat sits next to the glass, staring directly at the camera.
Shot 2, 2s: Close-up on the cat's paw slowly pushing the glass toward the edge.
Shot 3, 3s: Wide shot, static. The glass falls and shatters on the floor, milk splashing, the cat does not move.
Audio: glass sliding on the counter, a crash, a person off-screen gasping.

2. Horror reveal (4 shots, 12s)

A dark suburban bedroom at night, blue moonlight through the blinds. A boy of about ten in striped pajamas. Cinematic horror, cool desaturated color.
Shot 1, 3s: Wide shot, static. The boy lies awake in bed, eyes open.
Shot 2, 3s: Close-up of his face as he slowly turns his head toward the closet.
Shot 3, 4s: Slow dolly push toward the closet door, which is open a few centimeters, total darkness inside.
Shot 4, 2s: Extreme close-up of the boy pulling the blanket over his face.
Audio: a clock ticking, the closet door creaking, his breath catching. No music.

3. Product ad (4 shots, 10s)

A minimalist bathroom with white tile and soft window light. A green glass bottle of face serum with a dropper cap. Clean commercial look, shallow depth of field.
Shot 1, 3s: Wide shot, slow dolly push. The bottle sits on the edge of the sink.
Shot 2, 2s: Macro insert. A single drop falls from the dropper.
Shot 3, 3s: Close-up. A woman presses the serum into her cheek with her fingertips and smiles.
Shot 4, 2s: Static hero shot of the bottle with water droplets on the glass.
Audio: a soft drip, gentle room tone, no music.

4. Two-person dialogue (4 shots, 12s)

A small diner booth at night, neon sign glowing outside the window, rain on the glass. Maya, a woman in her twenties in a yellow raincoat. Leo, a man in his twenties in a denim jacket. Cinematic, warm and cool contrast.
Shot 1, 3s: Wide shot, static. Maya and Leo sit across from each other, coffee cups between them.
Shot 2, 3s: Over-the-shoulder on Maya. Maya (quiet, English): "I sold the car."
Shot 3, 3s: Close-up on Leo, his cup stopping halfway to his mouth. Leo (stunned, English): "You what?"
Shot 4, 3s: Close-up on Maya, trying not to smile.
Audio: rain on the window, diner chatter, a coffee machine hissing.

5. Travel montage (6 shots, 15s)

A solo female traveler with a green backpack in Lisbon on a sunny day. Warm cinematic travel look, golden light.
Shot 1, 3s: Wide aerial-style shot of red rooftops and the river.
Shot 2, 2s: Medium shot, tracking. She walks up a steep cobblestone street.
Shot 3, 2s: Insert. A yellow tram passes close to camera.
Shot 4, 3s: Close-up. She bites into a custard tart and laughs.
Shot 5, 2s: Low angle. She leans on a railing at a viewpoint.
Shot 6, 3s: Wide shot, slow dolly back. She looks out over the city at sunset.
Audio: tram bell, street guitar in the distance, seagulls.

For more destination ideas, the travel destination prompts page pairs well with this structure.

6. Recipe short (5 shots, 12s)

A rustic kitchen counter with a wooden board and a cast iron pan, soft window light. Only a cook's hands are visible. Top-down and close-up food video look.
Shot 1, 2s: Top-down shot. Garlic cloves are smashed with the flat of a knife.
Shot 2, 3s: Close-up. Butter melts and foams in the hot pan.
Shot 3, 2s: Insert. Garlic hits the butter.
Shot 4, 3s: Close-up, slow push. A steak is laid into the pan and basted with a spoon.
Shot 5, 2s: Top-down. The sliced steak on the board, pink inside.
Audio: knife thud, loud sizzle, spoon scraping the pan. No music.

7. Character vlog (3 shots, 10s)

A cozy small bedroom decorated with fairy lights. @Nova, a cheerful young woman with short pink hair in an oversized sweater. Vertical selfie vlog look, handheld phone camera.
Shot 1, 4s: Selfie medium close-up. @Nova (excited, English): "Okay, day three of learning guitar."
Shot 2, 3s: Close-up of her fingers pressing a chord badly on an acoustic guitar.
Shot 3, 3s: Selfie close-up. She winces and laughs at the camera.
Audio: a buzzing wrong chord, her laughter, quiet room tone.

Swap @Nova for your own Element. For more recurring-character ideas, see character vlog prompts.

8. Automatic Multi-Shot draft

When you just want to see how Kling would cut a scene, use automatic mode with a single descriptive prompt:

A street basketball game on an outdoor court at sunset, chain-link fence and city skyline behind. A skinny teenager in a white tank top steals the ball from a much taller player, sprints down the court and makes a layup while his friends go wild on the sideline. Energetic handheld sports footage, warm golden light. Audio: sneakers squeaking, ball bouncing, friends shouting.

If the cuts land in the right places, keep it. If not, copy the moments you liked into a Custom Multi-Shot prompt and set the durations yourself.

Mistakes That Break Multi-Shot Sequences

• Too many shots for the length. Start with two or three.
• No shared description. Without a line describing the characters and setting once, details drift between shots.
• Unrelated shots. Jumping from a kitchen to a beach with no connection gives you jarring cuts. Every shot should belong to the same scene.
• Overlong prompts. Kling-focused guides warn that prompts past roughly 200 to 250 words start to lose focus. Keep each shot to one or two short sentences.
• Several camera moves per shot. One move per shot. The cut is your second move.
• Untagged dialogue. Name the speaker every time.

Start With Good Scene Ideas

Multi-shot turns a good idea into a finished scene in one generation. It does not come up with the idea. Most of the prompts above started as a one-line concept, “cat knocks milk off counter,” and got broken into shots afterwards.

LzyPrompt is where that one-line concept comes from. Type an idea in plain words and it gives you 10 prompt variations, each with setting, lighting, action beats and sound cues, plus a caption ready to post. Pick the strongest one, split its action beats into shots, and paste it into Kling’s Custom Multi-Shot mode. The Kling AI prompt generator page shows the workflow, and our Kling 3.0 vs Veo 3.1 vs Sora 2 comparison explains when Kling is the better pick.

Try LzyPrompt free for 7 days and turn your next idea into a storyboard-ready scene.