Craft

Kling AI prompts for ads

By James Tumbucon
Updated · 6 min read

Kling AI prompts follow the order Kling documents: subject, action, setting, camera, lighting, then sound. Write each spoken line after a speaker label with the tone in parentheses, and set duration and aspect ratio in the settings. Below are original Kling AI prompts for seven ad formats in that structure.
On this page
  1. How do you write a Kling AI prompt?
  2. How do you write dialogue and sound in Kling?
  3. Kling prompt for a UGC selfie ad
  4. Kling prompt for a podcast clip ad
  5. Kling prompt for a street interview ad
  6. Kling prompt for a short drama ad
  7. Kling prompt for a claymation or animation ad
  8. Kling prompt for a product shot ad
  9. Kling prompt for a brand film ad
  10. How do you keep Kling characters consistent across clips?

How do you write a Kling AI prompt?

Write a Kling prompt like shot direction, in a fixed order. Kling's 4.0 guide gives the formula Subject + Action + Setting + Camera + Lighting + Pacing + Sound.

Kling's earlier text-to-video guide uses nearly the same order: subject and description, subject movement, scene, then optional camera language, lighting and atmosphere. It asks for simple words and short sentences, and warns that models are not sensitive to numbers and that complex physics, such as a bouncing ball, is hard to generate.

Element What to write for an ad Notes from Kling's guides
Subject The presenter or product, 2 to 3 visible traits Appearance can be a few short sentences
Action One clear action Keep it simple enough for the clip length
Setting Location and time of day Foreground and background
Camera Shot size, angle, one move Close-up, low angle, aerial and similar terms
Lighting Light source and mood Morning light, sunset, artificial light
Sound Dialogue, ambience, music Speaker label plus tone for each line

How do you write dialogue and sound in Kling?

Tie every line to a named speaker, with the delivery in parentheses. Kling's VIDEO 3.0 guide says that specifying dialogue for each character lets the model match each line to the right speaker.

The guide's examples follow the pattern Name (tone, accent or language): line, and a voiceover can be tagged the same way with a description of the voice. Ambient sound and background music are written as plain sentences. If a character element already has a bound voice, don't describe the voice tone again.

For image-to-video, Kling's image-to-video guide reduces the formula to subject plus movement, because the image sets the scene. Duration, aspect ratio and multi-shot mode are chosen in the settings.

The templates below are written for this page and have not been run through Kling; no outputs are shown. Square-bracket items are placeholders. For specs and workflow, see Kling for ads, or fill a template from a form in the AI ad prompt builder.

Kling prompt for a UGC selfie ad

A Kling UGC selfie prompt keeps one presenter, one action and one line per clip. See the get-ready UGC examples for the look.

A [age range] woman with [hair], wearing [top], holding [product].
She tilts the product toward the phone camera and taps the label twice.
Setting: a small bathroom, mirror behind her, early morning.
Camera: front-facing phone camera, eye level, slight handheld shake.
Lighting: soft daylight from a window on the left.
Woman (relaxed, slightly amused, American accent): [hook line]
Sound: a tap running quietly in the background, no music.

Kling prompt for a podcast clip ad

For a podcast clip, name both speakers and give each a tone tag so the lines land on the right mouth. Compare with the AI podcast ad examples.

Two podcast hosts at a wooden desk, each with a microphone and headphones.
Host A: a man in his [age] with [glasses / beard]. Host B: a woman with [hair].
Host A leans back; Host B picks up [product] from the desk.
Setting: a small studio with acoustic panels and a lamp.
Camera: static two-shot at eye level.
Lighting: warm practical lamp light, darker background.
Host A (curious, British accent): [question about the problem]
Host B (confident, steady pace): [answer that names the product]
Sound: close microphone sound, light room tone.

Kling prompt for a street interview ad

Street interviews work as one passer-by, one question and one answer per clip, with the interviewer off screen. See the AI street interview examples.

A [age range] man in a [jacket] on a city pavement, holding a coffee cup.
A handheld microphone enters from the right edge of frame.
He stops walking, looks at the microphone and grins.
Setting: busy shopping street at midday, people passing behind him.
Camera: handheld medium shot, slightly low angle.
Lighting: bright overcast daylight.
Interviewer (off camera, upbeat): [quick question]
Man (laughing, casual): [short answer with the product name]
Sound: street noise, footsteps, a distant car horn.

Kling prompt for a short drama ad

Short drama needs the most dialogue, so this is where speaker tags matter. Keep it to two characters and one turn. See the AI drama ad examples.

Shot 1: A woman in [work clothes] stands at a kitchen counter, staring at [problem object].
Her husband in a [jumper] walks in behind her.
Camera: static medium shot. Lighting: cool evening light from the window.
Husband (gentle, concerned): [question about the problem]

Shot 2: Close-up of the woman. She exhales and rubs her forehead.
Woman (tired, quiet): [line admitting the problem]

Shot 3: The husband sets [product] on the counter. She picks it up and smiles.
Camera: slow push-in. Lighting: warmer.
Woman (surprised, brighter): [payoff line]
Sound: soft piano music begins under the last line.

Use the shot labels with Multi-Shot switched on in Kling 3.0; the guide notes the model may still render one shot if it judges that better suits the scene.

Kling prompt for a claymation or animation ad

For stylised formats, put the style first so the model does not drift toward live action. See the claymation and craft examples and the Pixar-style examples.

Claymation style, handmade clay figures and props.
A small clay [product] character with round eyes and stubby arms.
It hops out of [box / drawer], lands on the edge, and wobbles.
Setting: a clay [kitchen / desk] scene in miniature scale.
Camera: low angle close to the surface, slow push-in.
Lighting: soft warm studio light, gentle shadows.
Clay character (squeaky, cheerful): [one-line benefit]
Sound: soft clay squish on landing, light playful music.

Kling prompt for a product shot ad

A product shot has no speaker, so the sound line carries the effect. See the AI product demo examples.

A [product] in [colour and material] standing upright on [surface].
Water drops slide down its side and pool at the base.
Setting: clean studio set, [colour] background.
Camera: macro close-up, slow orbit to the right, ending on the label.
Lighting: hard rim light from behind, soft fill from the front.
Sound: a light splash, then quiet.

For image-to-video, drop the first and third lines and describe only the movement: the drops, the orbit and the light change.

Kling prompt for a brand film ad

Brand films lean on setting, light and sound, with no lines. Kling 4.0 adds a 21:9 ultrawide format for this kind of shot. See the AI brand film examples.

A [age range] man in a [coat], carrying [product], walking alone.
He stops at the top of a hill and looks out over the valley.
Setting: [mountain road / harbour] at sunrise, mist low on the ground.
Camera: wide aerial shot, slow drift forward, then a slow descent.
Lighting: low golden sun, long shadows.
Sound: wind, distant birds, a slow cello line.

How do you keep Kling characters consistent across clips?

Bind the presenter as an element and reuse it for every clip of the ad. Kling's 3.0 guide says elements can come from a character video or from 2 to 4 reference images, and the 4.0 guide says to give each reference asset a role: character, product, scene, style or video motion.

For one script across every format, read one script, every format. If your Kling ads already perform on paid traffic, see how Anglr works for creators.

Who this is for

  • Independent AI creators and editors. You make the ads yourself. You know your work performs, so you'd rather be paid on what it does than on a flat fee.
  • AI agencies and studios. You run a team of editors. You write the brief, your team makes the versions, and you're paid on performance because you know your team's ads scale.

Not for you if

  • You only want a flat fee.
  • You're new to editing and not yet confident your ads will perform.
  • You want every shot spelled out for you.
  • You make one style and won't try others.

FAQ

What is the Kling prompt formula?

Kling's 4.0 guide gives Subject + Action + Setting + Camera + Lighting + Pacing + Sound. Its older text-to-video guide gives Subject + Subject movement + Scene, with camera language, lighting and atmosphere as optional extras.

How do you write dialogue in a Kling prompt?

Write the character's name, the tone or accent in parentheses, a colon, then the line. Kling's 3.0 guide says pairing each line with its speaker lets the model match speech to the right character, even with three or more people.

What languages can Kling characters speak?

Kling's 3.0 guide lists Chinese, English, Japanese, Korean and Spanish, and says dialogue in other languages is translated into English. Characters can switch language inside one video.

What should a Kling image-to-video prompt include?

The subject and how it moves. The image already sets the scene, and Kling's guide says a description that differs a lot from the image can cause a cut or transition.

How long can a Kling prompt be?

Kling's 4.0 guide says prompts can run up to 8,000 tokens. Its earlier guides still recommend simple words and sentence structures, so length is no reason to add detail the shot does not need.

Sources

Checked: 12 Oct 2026