How to write prompts for Kling V3
Kling V3, which Kling’s guide calls VIDEO 3.0, turns words or a start picture into a video, with sound. The first two sections put Kling’s own guide in our words, and we did not test them. The last rests on six sample clips made with this model that we watched ourselves, a small sample.
Written from the provider’s documentation and from results we checked ourselves. Last checked November 8, 2026.
Write it like a shot list
From the provider’s documentation
- Kling’s guide asks for detailed, shot-by-shot writing: the setting first, then the subject, the actions, the camera moves and the speech, in the order they happen. Its examples use cinematic camera words such as tracking, orbit, pan, zoom in and handheld.
- Keep physical actions and the feeling of each beat explicit. For a long clip, its example marks time in words, such as “at the 4th second” and “in the final 3 seconds”, so the model can follow the sequence.
- For one unbroken shot, Kling’s own examples say so in the prompt, with wording such as “one continuous long take without any cuts” and “a single unbroken shot with no edited transitions”.
- Kling’s own screen has a Multi-Shot switch. With it on, the model plans the cuts, angles and framing itself. With it off, it makes a single shot. A Custom Multi-Shot option lets you set the number and length of the shots, labelled Shot 1, Shot 2 and so on. Pickshot’s form has no such switch, so put the shots, or the single take, in your words.
Speech, languages and start pictures
From the provider’s documentation
- To make characters speak, pair each speaker with their exact line. Kling says VIDEO 3.0 handles three or more speaking characters better than its predecessor VIDEO 2.6. Add a tone to a line, such as “softly, in a surprised tone”.
- Kling lists five languages for speech: Chinese, English, Japanese, Korean and Spanish, with switching between them inside one scene. Speech in any other language is translated into English. Name a dialect or an accent next to the line, such as Cantonese or Indian-accented English.
- For a start picture, Kling’s guide shows a prompt that takes what the picture already holds as given and then says what the camera and the subject do next. It says signs, captions and logos in the start picture are kept without drifting or blurring. We did not test this.
- Kling’s VIDEO 3.0 guide does not discuss negative prompts, so describe what you want to see and hear.
What our sample clips showed
From results we checked ourselves
- In the bamboo clip two duellists, one in white robes and one in black, fight with swords among the bamboo, with motion blur on the swings. They are small in the wide frame, and we did not see the leaves burst up around them as the prompt said. They lie on the ground.
- In the snowball clip a shiba inu leaps and catches a snowball in its mouth in a snowy yard. It wears a red scarf that the prompt did not mention.
- The candle clip shows a hand with a lighter working along a long row of candles on a laid dining table, and more candles are lit as it goes, which matches “one by one”.
- The Shinjuku alley prompt asked for glowing signs without text. The neon signs show shapes and marks and no readable words. The camera moves forward through the alley and the ground is wet, although we could not see rain falling.
- The dragon clip shows a green veined egg cracking and a 3D-animated hatchling looking out. Its eyes close for a moment around the four-second mark, which fits the blink in the prompt. The Jeju clip shows a woman in a flowing linen dress walking along a black stone wall under a sunny sky, with wind in her dress and her hair, filmed from behind.
- Each prompt was a single sentence, so these clips say nothing about numbered shots or speech. We did not judge the sound of any clip.
Real results and the prompts behind them
Results made with this model. The prompts are shown exactly as written, in English.
Settings of this model
- Od
- Kuaishou
- Typ
- AI video model
- Způsob tvorby
- Text na video · Z obrázku video
- Délka
- 3–15s
- Rozlišení
- 720p / 1080p / 4K
- Aspect ratios
- 16:9 · 9:16 · 1:1 · 4:3 · 3:4
- Zvuk
- Se zvukem
- Cena
- from 142 / sec
Otázky
How do I keep Kling V3 to one shot?
Kling’s own examples say it in the prompt: “one continuous long take without any cuts”. Its screen has a Multi-Shot switch that Pickshot’s form does not show. We did not test how reliably the words keep a single shot, so check the result.
Can Kling V3 speak Korean or Japanese?
Kling lists Chinese, English, Japanese, Korean and Spanish for speech and says other languages are translated into English. We did not judge speech in our samples.
Will signs I ask to be blank stay blank?
In our Shinjuku sample the neon signs show shapes and no readable words. One clip is not enough to promise it, so check the result when it matters.
Provider documentation we read
We read these pages and wrote the points that matter in our own words. The provider may change them.
- Kling VIDEO 3.0 model user guide (Kling) (kling.ai)
More prompting guides
- How to write prompts for GPT Image 2.6 FlareAI image model
- How to write prompts for GPT Image 2.6 SunburstAI image model
- How to write prompts for Nano Banana 2.1AI image model
- How to write prompts for MOSS ImageAI image model
- How to write prompts for Grok Imagine ProAI image model
- How to write prompts for Nano Banana ProAI image model
- How to write prompts for Nano Banana 2AI image model
- How to write prompts for GPT Image 2.5 FlareAI image model
- How to write prompts for IndigoAI image model
- How to write prompts for SaffronAI image model
- How to write prompts for CeladonAI image model
- How to write prompts for LapisAI image model
- How to write edit requests for UmberAI editor obrázků
- How to write prompts for Kling Video 4.0AI video model
- How to write prompts for Seedance 2.5AI video model
- How to write prompts for Seedance 2.0 MiniAI video model
- How to write prompts for Kling O3 OmniAI video model
- How to write prompts for Veo 3.1AI video model
- How to write prompts for Veo 3.1 FastAI video model
- How to write prompts for SilkAI video model
- How to write prompts for LacquerAI video model
- How to write prompts for SatinAI video model
Názvy modelů, společností a produktů na tomto webu jsou ochrannými známkami jejich vlastníků; Pickshot s nimi není spojen, sponzorován ani jimi schválen.