Kling 4.0 Text to Video

Kling 4.0 Text to Video — AI Text-to-Video Generation

What is Kling 4.0 Text to Video?

Kling 4.0 Text to Video is Kuaishou's flagship text-to-video model, turning a written prompt into a 3–30 second clip with native audio — no source image or reference footage required. It is the text-only entry point to the Kling 4.0 family, built to fix what earlier AI video found hardest: holding fast, large-scale motion together while the camera moves. High-speed action and sweeping camera work — push-ins, pull-outs, pans, tracking and orbit shots — stay stable and continuous, so the motion reads cleanly instead of warping. Output runs at 720p or 1080p in 16:9, 9:16, 1:1 or 21:9, and sound is generated in sync with the picture.

Key Features

• •Native stereo audio: two-channel dialogue, sound effects and ambience rendered with the video, with improved lip sync.
• •Stable large-scale motion and camera control: tracking, orbit and push-in shots stay coherent through high-speed action.
• •Single-pass duration 3–30 seconds: long takes and continuous storytelling without stitching clips.
• •Four aspect ratios: 16:9, 9:16, 1:1 and 21:9 for cinema, vertical social, square and ultrawide.
• •Prompts up to 26,000 characters covering subject, motion, camera direction and sound in one block.
• •Multilingual dialogue across many languages, with legible on-screen text.

Best Use Cases

Kling 4.0 shines where motion is the point: sport and dance, machinery and vehicles, crowds and live performance. In testing, a prompt for a b-boy's windmill-into-freeze returned a clean 1080p clip where the body stayed coherent through the spin while the camera orbited, carrying a synced stereo mix of beat, sneaker scuffs and breath. Because audio is native, the model suits dialogue-led scenes, UGC and character ads, where speech and sound are generated with the shot. The 30-second single-pass ceiling supports a full short-form ad with several beats, a continuous reunion shot, or a talking vlog.

Prompt Tips and Output Quality

Lead with the subject and one clear action, then name the camera move that the action motivates, then the sound you want. Keep negatives inside the prompt — there is no separate negative field. Avoid piling on many simultaneous actions or competing camera directions; a dense brief is harder to honor and can invent details, so review every clip. Set duration and aspect ratio explicitly rather than only describing them. On this integration the output ceiling is 1080p.

How to Use Kling 4.0 Text to Video

Kling 4.0 Text to Video builds a complete clip from a single prompt, so the prompt is where almost all of your control lives. Write it as a short scene, not a keyword list: name the subject, give it one clear action, state the camera move that the action motivates, then describe the sound. Because audio is generated natively, cues like "a thumping beat, footsteps on gravel, distant traffic" or a line of spoken dialogue in quotes become part of the output.

Writing effective prompts

Put everything in one block. Start with who or what and the single action you want to see — a windmill into a freeze, a kettle pouring, a cyclist cresting a hill. Follow with the camera: orbit, tracking shot, push-in, pan. Add one lighting note and the ambient or spoken sound. Keep any negatives ("no on-screen text, no cuts") in the same prompt, since there is no separate negative field. Resist stacking several simultaneous actions or competing camera directions — a focused brief is honored more faithfully, and every clip still deserves a review pass.

Choosing parameters

• •duration: about 5 seconds for a single action beat; 10–30 seconds for dialogue or a multi-beat story such as a short ad.
• •resolution: 1080p for hero and final deliverables; 720p for faster drafts and iteration.
• •aspect_ratio: 16:9 for landscape, 9:16 for vertical social, 1:1 for square, 21:9 for ultrawide. Set it explicitly rather than only describing framing in prose.

Recommended Settings

The tested default — prompt describing a b-boy's windmill-into-freeze with an orbiting camera, duration 5, resolution 1080p, aspect_ratio 16:9 — returned a clean 1920×1080, 24 fps clip with a two-channel stereo audio track: fast motion stayed coherent through the spin and the camera orbit held. For dialogue-led or UGC shots, raise duration toward 10 seconds, keep the frame uncluttered so speech lands, and write the spoken line in quotes. For motion showcases, favor one strong action plus a single motivated camera move.

Examples to try

Each example ran end to end at the settings shown and came back as a finished clip with its own audio; clicking a card drops its prompt and parameters back into the form. What carries over is the length, the action and camera move, the aspect ratio and the native stereo sound.

• •Large-scale motion with an orbiting camera (the default; 1080p, 16:9). Prompt: "A b-boy spins into an explosive windmill on a concrete plaza… then snaps into a one-handed freeze. The camera orbits him low and fast, keeping pace with the rotation… Sound: a thumping hip-hop beat, sneaker scuffs on concrete, and his sharp exhale as he hits the freeze." The body holds together through the spin while the camera keeps pace, over a synced stereo mix of beat, scuffs and breath. Tip: one strong action plus a single motivated camera move; ~5s per beat.
• •Spoken dialogue with lip sync (720p, 9:16). Prompt: "A marine biologist in a blue lab coat holds up a glass jar of glowing bioluminescent plankton and turns it toward the lens… She says, clearly: "Watch what happens when I shake it - the whole jar lights up." Sound: her bright voice close to the mic, a low filtration hum, and water bubbling in tanks…" Her mouth stays in sync with the line and the jar lights up on cue, voice foregrounded in stereo. Tip: put the spoken line in quotes and keep the frame tight so the words land.
• •A long single-pass continuous take (duration 10, 21:9). Prompt: "A single unbroken shot travels alongside a green combine harvester as it cuts a wide swath through a field of ripe wheat… The camera tracks level with the header, then rises on a crane move to look down the uncut rows ahead… Sound: the deep diesel engine, the thresher's steady roar, grain hissing into the trailer…" One unbroken take tracks the machine and rises on the crane move with no cut, grain and dust coherent throughout. Tip: raise duration for continuous action and name one sustained camera move.
• •Multilingual dialogue and legible on-screen text (1:1). Prompt: "A 1930s radio announcer in a pinstripe suit leans toward a chrome ribbon microphone and reads the evening news… a lit glass "EN DIRECT" lamp and a painted "RADIO PARIS" placard stay sharp and readable through the move… He says in French: "Mesdames et messieurs, voici les nouvelles de ce soir."" Both signs stay readable as the camera pushes in while he reads the French line. Tip: spell the exact on-screen words and name the spoken language in the prompt.

Also try: “a chef plates a dish and explains the recipe in Japanese”, “a two-person argument in one 15-second take”, or “a product hero spin with a tracking camera and a tagline read aloud”.

Does Kling 4.0 Text to Video generate sound?

Yes — native two-channel stereo audio, including dialogue, sound effects and ambience, timed to the picture.

How long can a clip be?

3 to 30 seconds in a single generation, set with the duration parameter.

What resolutions and aspect ratios are supported?

720p or 1080p, in 16:9, 9:16, 1:1 or 21:9.

Do I need an input image?

No. It generates the whole scene from text alone; the image-to-video and reference-to-video siblings handle media inputs.

Can it produce vertical video for social?

Yes — choose the 9:16 aspect ratio for feed and stories.

How do I get legible dialogue?

Write the spoken lines in the prompt and keep the scene uncluttered so the words land clearly.