Kling prompt generator
Kling’s guide builds a prompt from subject, movement and scene, then adds camera language, lighting and atmosphere if you need them. This generator follows that order and writes multi-shot prompts in the format the Kling 3.0 API reads.
Last updated: 2026-10-03
Kling’s prompt formula
Prompt = subject (with description) + subject movement + scene (with description) + optional camera language, lighting and atmosphere. Kling treats subject, movement and scene as the core. For image-to-video the image already sets the scene, so the guide says to describe only how the subject and the background move, in simple words and with physically plausible motion.
Here is the sample scene as a two-shot Kling 3.0 prompt:
shot 1, 3, A young woman in a red raincoat stops and looks up as the rain starts in a neon-lit Tokyo side street at night. Push in. Soft neon lighting. Quiet, nostalgic atmosphere, cinematic 35mm film style. Wide shot, she walks toward the camera under the neon signs; shot 2, 3, Close-up, raindrops land on her face and she smiles. The woman: "It finally came." Sound: rain on the pavement, distant traffic. No subtitles or on-screen text. No background music;Multi-shot syntax
- Write each shot as “shot n, m, words;” where n is the shot number, m its length in seconds and words its description, separated by standard semicolons.
- Use 1 to 6 shots, each at least 1 second. The seconds must add up to the clip’s total duration, which is 3–15 seconds on Kling 3.0.
- Each shot’s description can be up to 512 characters.
- Since May 7, 2026, multi-shot can be combined with first and last frames on kling-v3 and kling-v3-omni.
- The generator numbers the shots and adds up the seconds. It puts your overall description into shot 1 and the dialogue and exclusions into the last shot, warns if the total falls outside 3–15 seconds or a shot passes 512 characters, and turns semicolons in your text into commas so they don’t split a shot.
Camera language is not camera control
Kling’s guide separates camera language (shot size, lens, angle, depth of field: ultra-wide angle, bokeh, close-up, telephoto, low-angle, high-angle, aerial view) from camera motion control. The current API has no camera_control parameter, so moves such as a tracking shot, an orbit or a push in go into the prompt as words.
Dialogue and native audio
Kling 3.0 matches each character to their line when you name the speaker, as in Kling’s own example: Mom (softly, in a surprised tone): “…”. It speaks Chinese, English, Japanese, Korean and Spanish; lines in other languages are translated into English. Accent and dialect tags such as “in Cantonese” or “with an Indian accent” work too.
Negative prompts in Kling
The current API, which Kling 3.0, 3.0 Omni, 3.0 Turbo, 2.6 and 2.5 Turbo moved to on July 15, 2026, has no negative_prompt field: write what you don’t want as sentences inside the prompt. The legacy API still accepts negative_prompt (up to 2,500 characters), but Kling recommends negative sentences inside the prompt either way.
Kling 3.0 at a glance
| Kling 3.0 | |
|---|---|
| Clip length | 3–15 seconds |
| Resolution | 720p, 1080p, 4K |
| Aspect ratio | 16:9, 9:16, 1:1 |
| Prompt length | 3,072 characters (Kling recommends staying under 2,500); 512 characters per shot in multi-shot mode. |
| API model | kling-3.0 / kling-3.0-omni / kling-3.0-turbo (legacy API: kling-v3, kling-v3-omni) |
| Where to use it | Kling AI (kling.ai), 可灵AI in China |
Kling has announced Kling 4.0 for October 2026. It isn’t in the API yet; its product page lists 21:9 and a “Smart” aspect ratio.
Frequently asked questions
Does the Kling prompt generator work for older Kling models?
The subject, movement and scene structure comes from Kling’s general prompt guide and applies to every model. The “shot n, m, words;” syntax is for Kling 3.0, 3.0 Turbo and 3.0 Omni; Kling 2.6 and 2.5 Turbo prompts are capped at 2,500 characters.
How long can a Kling prompt be?
3,072 characters for Kling 3.0, 3.0 Omni and 3.0 Turbo text-to-video; Kling recommends staying under 2,500. Kling 2.6, 2.5 Turbo and the legacy API allow 2,500. In multi-shot mode each shot is limited to 512 characters.
How do I make a multi-shot video in Kling 3.0?
Write one “shot n, m, words;” entry per shot, up to six, where m is the shot length in seconds. The lengths must add up to the clip duration you set, between 3 and 15 seconds.
Can I use a negative prompt in Kling 3.0?
Not as a separate field in the current API. Put negative sentences inside the prompt, which is what Kling recommends even on the legacy API.
How do I reference an image in a Kling prompt?
In the current API, write @image_1, @video_1 or an element name such as @Zhang. Avoid element names that contain each other, such as @Zhang and @ZhangSan. The legacy API used <<<image_1>>>.
How much does Kling 3.0 cost per second?
$0.084 at 720p and $0.112 at 1080p without audio at Kling’s list price; native audio raises that to $0.126 and $0.168, and 4K is $0.42. The official API needs a prepaid package.
Sources: Kling text-to-video prompt guide · Kling image-to-video guide · Kling VIDEO 3.0 user guide · Kling 3.0 API reference
Related tools
- Kling 3.0 price per secondKling 3.0 price per second: $0.084 at 720p and $0.112 at 1080p without audio, $0.42 in 4K. See what audio adds and which platforms skip the $700 prepaid pack.
- Seedance prompt generatorFree Seedance prompt generator. Builds Seedance 2.5 prompts with timestamps and ByteDance’s sound tags, or Seedance 2.0 prompts in Shot 1 / Shot 2 format.
- Wan 3.0 prompt generatorBuild Wan 3.0 prompts in Alibaba’s recommended order: overall description, timed shots, dialogue, sound, style and a negative list. Free, no sign-up.
- MiniMax H3 prompt generatorWrite MiniMax H3 prompts the way the official cookbook lays them out: one core sentence, a shot-by-shot timeline and exclusions at the end. Works for Hailuo AI.