Skip to content

Kling 3.0 prompt generator

Kling 3.0 makes one continuous take of 3 to 15 seconds and can lip-sync a spoken line to the person who says it. This guide shows how a Kling prompt is built and what to avoid, and its generator writes one: a single unbroken shot built around the action and the camera move and, if someone speaks, the line with the language it is spoken in.

  • 11 credits a prompt (about $0.10 on Starter), with any DreamBench plan.
  • No photo needed, and no model is called.
  • A preset of the AI video prompt generator.
Go to the generator

What is a Kling 3.0 prompt generator?

It writes a Kling 3.0 prompt from your choices: one 3- to 15-second take built around the action and one camera move, with an optional line lip-synced in one of five languages, asking only for what Kling makes on DreamBench. It is part of every DreamBench plan, from $9.99 a month.

How to write a Kling 3.0 prompt

  1. Describe the action

    Who or what is in frame, what they do and where. Kling 3.0 is built for movement, so give the subject something to do: a sprint, a spin, a pour, a skid.

  2. Direct the take

    3, 5, 10 or 15 seconds in one continuous take, the frame, the light, the lens and the one move the camera makes. For dialogue, choose One line of dialogue, type the line and pick the language it is spoken in.

  3. Copy or run it

    The prompt opens with the frame and the take, puts the camera move with the action and keeps the subject steady through it. The first time, create a free account and choose a plan, then copy it into any tool that runs Kling 3.0, or run it on DreamBench.

Start from an example

Fill in the subject and what they do, and the prompt appears here.

A prompt uses 11 credits from your plan, and the same prompt again uses none. Running it on DreamBench renders a real clip, priced on its own.

How a Kling 3.0 prompt is built

Tap a part to see where it lands in a real prompt, or tap a sentence of the prompt to see which part wrote it.

  1. The frame and the take

    16:9, 9:16, 1:1, 4:3 or 21:9, then one continuous take of 3, 5, 10 or 15 seconds with no cuts, said first.

  2. The action, with its line

    Who is in frame and what they do, in motion: Kling 3.0 is built for movement. A spoken line goes in quotes after the person who says it, with the language it is said in.

  3. The camera with the action

    One shot size, a lens and one move, straight after the action, because the move is part of what happens.

  4. The subject held through the move

    A sentence that keeps the subject and everything around it in shape as the camera moves, which is what Kling designed 3.0 to do.

  5. The light and the look

    The source and its side, then one register, such as photoreal film or black-and-white.

  6. The sound last

    The scene's ambience, or the line in its language with the lips in time to every word; or silence.

A Kling 3.0 prompt

Widescreen 16:9, one continuous 5-second take with no cuts. A sprinter in a red tank top explodes out of the blocks and races down the track, on an empty athletics track at sunset. Wide shot on a 35mm lens at a few metres: the subject and the room both readable, little distortion. One move only: a lateral track that keeps pace with the subject, passing a foreground shape. The subject and everything around it keep their shape and look as the camera moves. Light: low golden-hour sun from behind the subject, the camera shooting into the light from the shadow side, a rim of light on the edges; the exposure holds the subject; no fill from the camera side. Photoreal, shot on film: natural colour, gentle grain, true skin. Sound: the scene's own ambience, no music. No text, no captions, no watermark.

Example prompts

Each one is written by the generator on this page, from the choices its note describes.

A sprint in one take

Kling 3.0

Five seconds of fast action with one lateral track keeping pace, and a sentence that holds the runner's look through the move.

Prompt

Widescreen 16:9, one continuous 5-second take with no cuts. A sprinter in a red tank top explodes out of the blocks and races down the track, on an empty athletics track at sunset. Wide shot on a 35mm lens at a few metres: the subject and the room both readable, little distortion. One move only: a lateral track that keeps pace with the subject, passing a foreground shape. The subject and everything around it keep their shape and look as the camera moves. Light: low golden-hour sun from behind the subject, the camera shooting into the light from the shadow side, a rim of light on the edges; the exposure holds the subject; no fill from the camera side. Photoreal, shot on film: natural colour, gentle grain, true skin. Sound: the scene's own ambience, no music. No text, no captions, no watermark.

A line in Japanese

Kling 3.0

A line typed in Japanese, with the language named, so Kling voices it and times the lips to it, in a vertical frame for a phone.

Prompt

Vertical 9:16, one continuous 10-second take with no cuts. A young chef in a white jacket slides a bowl of ramen across the counter and smiles, in a narrow ramen bar at night, and says in Japanese, "どうぞ、熱いうちに召し上がってください。" Medium shot on a 50mm lens at a couple of metres: natural proportions, the background softening behind the subject. One move only: a slow push-in that ends close on the subject. The subject and everything around it keep their shape and look as the camera moves. Light: night, lit only by the practical lights in the scene from camera-left, so the subject's camera-right side falls to shadow; the exposure holds the subject; no fill from the camera side. Photoreal, shot on film: natural colour, gentle grain, true skin. Sound: the line in Japanese, the lips in time with every word, over the scene's own ambience. No text, no captions, no watermark.

An orbit around a car

Kling 3.0

A slow orbit, the kind of move Kling 3.0 is built to hold steady through, in a cinematic 21:9 frame.

Prompt

Cinematic 21:9, one continuous 10-second take with no cuts. A vintage red sports car idles on a wet cobbled street, its headlights cutting through the drizzle, in an old town square at night. Medium shot on a 35mm lens at a few metres: the subject and the room both readable, little distortion. One move only: a slow orbit in checkpoints, front, then the far side, then back. The subject and everything around it keep their shape and look as the camera moves. Light: night, lit only by the practical lights in the scene from behind the subject, the camera shooting into the light from the shadow side, a rim of light on the edges; the exposure holds the subject; no fill from the camera side. Photoreal, shot on film: natural colour, gentle grain, true skin. Sound: the scene's own ambience, no music. No text, no captions, no watermark.

Common mistakes

A subject with nothing to do

Kling 3.0 is built for movement, and a subject standing still for 15 seconds wastes it. Give them an action: a sprint, a spin, a pour.

Two moves at once

A track that turns into an orbit smears. Pick one move and pair it with the action.

Cutting to a new place

On DreamBench a Kling clip is one continuous take and doesn't cut to a new setting partway through. A second place is a second clip.

A line in a sixth language

Kling lists Chinese, English, Japanese, Korean and Spanish for dialogue, and says it translates lines in other languages into English. Keep to its five.

Leaving out the accent

Kling can give a voice an accent, such as British or Indian English, or a Chinese dialect such as Cantonese, if the prompt asks. Put it in the subject: an old man with a British accent.

Works with

What it takes
Your own words: who is in the shot, what they do and where. No photo is needed.
What you get
A prompt as plain text.
DreamBench's composer
Use it as the brief of a film on Kling 3.0: DreamBench plans the shots and draws each one’s opening frame, at 720p on Kling 3.0 Standard or 1080p on Kling 3.0 Pro.
Kling's own app
The same plain-text prompt runs there, where a multi-shot mode DreamBench doesn't use is also offered.
The Kling API
Send the prompt as text, the line and its language included.

Limits

Which model it writes for, and its limits

Writes for Kling 3.0, within what it can make on DreamBench:

Lengths
3, 5, 10 or 15 seconds
Frames
16:9, 9:16, 1:1, 4:3 or 21:9
Sound
On or off: ambience or a spoken line
Shots
One continuous take, no cuts
Dialogue
A line lip-synced in English, Chinese, Japanese, Korean or Spanish
Prompt format
Plain text

One continuous take

Every prompt is one unbroken shot of 3 to 15 seconds, the way Kling 3.0 runs on DreamBench, with no cut to a new place partway through.

Lip-synced lines in five languages

A spoken line is written with the language it is in: Chinese, English, Japanese, Korean or Spanish, the five Kling lists, and Kling matches the voice and the lips to the person who says it.

Steady through the move

Kling designed 3.0 to keep characters and objects consistent while the camera tracks, swings or circles, so the prompt pairs one clear move with a sentence that holds the subject's look.

Every frame it makes here

Widescreen 16:9, vertical 9:16, square 1:1, 4:3 or cinematic 21:9, the frames Kling 3.0 renders on DreamBench, named first in the prompt.

Where it fails

  • On DreamBench a clip is one continuous take of up to 15 seconds and doesn't cut to a new setting.
  • Dialogue is lip-synced in five languages; Kling translates a line in any other language into English.

Kling 3.0 prompt generator FAQ

How do I write a prompt for Kling 3.0?

To write a prompt for Kling 3.0, open with the frame and one continuous take of 3, 5, 10 or 15 seconds, then who is in it, what they do and where, the shot size, a lens and one camera move, the light and the sound. Give the subject a clear action: movement is what Kling 3.0 is built for. The generator writes it in that order from your choices.

How does Kling 3.0 lip-sync dialogue?

With sound on, Kling 3.0 gives each speaking character their own voice and times the lips to their lines. The generator writes your line in quotes after the subject, with the language it is spoken in. Kling lists five languages: Chinese, English, Japanese, Korean and Spanish, and says it translates lines in other languages into English.

How long can a Kling 3.0 clip be?

From 3 to 15 seconds of continuous video in one generation; the generator offers 3, 5, 10 or 15. On DreamBench a clip doesn't cut to a new setting partway through, so a longer story is several clips, which the composer plans and joins. For a square or 21:9 frame Kling and Seedance both serve; Veo makes only 16:9 and 9:16.

Where is Kling's AI video prompt generator, and is this one the same?

Inside Kling's video generator: the DeepSeek icon at the bottom right of the prompt box returns three suggested prompts, which Kling's own article says describe the scene but don't write the motion or the camera. This generator writes both, and needs a DreamBench plan: you create an account on the way, and the button shows the price in credits before you press it.

How do I write a Kling image to video prompt?

A Kling image to video prompt describes only what changes. The image already fixes the subject, the layout and any text, so say how the scene evolves: one movement, one camera move, a shift in the light, and name the subject the same way in every clip. This generator writes text-to-video prompts; on DreamBench the logo animator runs Kling 3.0 from your picture with a first and a last frame.

Should I use Kling's negative prompt?

This generator writes none: it closes with one line, no text, no captions, no watermark, and puts everything else as a positive direction, which video models follow more reliably than a list of things to avoid. If the tool you paste into offers a separate negative-prompt box, keep it to a few short nouns, such as blur or extra fingers.

How should I write a Kling multi-shot prompt?

On DreamBench, Kling 3.0 runs as one continuous take, so this generator writes one shot. In Kling's own app, its user guide describes a Multi-Shot switch where you set the number of shots and the length of each. For numbered, timed shots in one clip here, use the Seedance 2.5 prompt generator.

How do I write two people talking in Kling?

This generator writes one line for one speaker. For an exchange, write it by hand: name each speaker the same way every time, put each one's action before their line, and keep to Kling's five languages. Or make two clips, one line each, and join them in the composer.

Can I write the Kling prompt in Chinese?

The generator writes its own sentences in English and passes your words on as typed, so a line of dialogue in Chinese goes in as it is, and Kling lip-syncs it. Kling's own guide writes its example prompts in English with the Chinese, Japanese or Korean line inserted, so type the subject, action and place in English for the most reliable result.

Does a Kling 2.6 prompt work on Kling 3.0?

Yes: the structure carries over, and Kling 3.0 is the successor to Kling 2.6. Write for what 3.0 adds, a take of up to 15 seconds and lines lip-synced in five languages; the generator offers only what Kling 3.0 makes on DreamBench.

Does the prompt change for Kling 3.0 Pro?

No. The same prompt runs on Kling 3.0 Standard at 720p and on Kling 3.0 Pro at 1080p in the DreamBench composer; the tier sets the resolution, not the prompt. Write the take, the action, the move, the light and the sound once, and pick the tier where you run it.

Can I ask for an accent?

Yes. Kling says you can ask for an accent, such as British or Indian English, or a Chinese dialect such as Cantonese. Put it in the subject, for example an old man with a British accent, and the generator carries it into the prompt with the line.

Can a Kling 3.0 clip cut between shots?

Not on DreamBench: Kling 3.0 runs here as one continuous take, and a clip doesn't cut to a new setting partway through. Kling's own app has a multi-shot mode; for several linked shots in one clip, the Seedance 2.5 prompt generator writes them.

How much does the Kling 3.0 prompt generator cost?

Each prompt uses 11 credits from your DreamBench plan, and the same prompt again uses none. Starter, $9.99 a month, includes 1,100 credits a month, about 100 prompts, plus AI video and the image tools.

Do I need an account to use the Kling 3.0 prompt generator?

Yes: a free account and a DreamBench plan. The tools are part of every plan, from Starter at $9.99 a month.

Copy checked against the engine on .

More tools

Every tool

Learn

Guides, comparisons and terms

Then make it move

Every result can go into a video.

Run your prompt in the composer, and DreamBench writes the scenes and renders them on the model you pick, from Seedance 2.5 to Veo 3.1 and Kling 3.0. Rendering uses credits, which come with a plan.

See the plans