A curated library of prompts for Kuaishou's Kling v3, organized by use case: cinematic single shots, multi-shot storyboards, lip-synced dialogue, and stylized sequences. Copy any prompt, adapt the details, and run it in VideoGen.
Kling v3 generates up to 15 seconds in one take, so a single well-described camera move can breathe. These prompts give each shot one subject, one move, and one mood.
Slow push-in on an old watchmaker at his bench, magnifying loupe over one eye, tiny gears glinting in warm lamplight, dust motes drifting, quiet ticking filling the room
Fifteen seconds rewards patience. One slow camera move with rich micro-detail beats three rushed beats.
Wide tracking shot alongside a night train crossing a snowy plain, warm windows streaking past, moonlit steam trailing, low rumble and rhythmic rail clicks
Lateral tracking shots with a repeating rhythm (windows, rail clicks) hold visual interest across long clips.
Overhead crane shot descending into a night market: lantern strings, steam from food stalls, crowds parting around a cyclist, sizzling woks and layered chatter
A descending crane move gives natural narrative shape: establish wide, end close on human detail.
Slow-motion shot of a matador's cape swirling in golden arena light, fabric unfurling in heavy folds, dust kicked up glowing at sunset, dramatic Spanish guitar
Fabric physics in slow motion is a Kling strength. Describe the weight of the material, not just its color.
Kling v3 supports multi-shot storyboarding with up to 6 camera cuts in one generation. Number the beats and give each shot one setup.
Four-shot sequence: close-up of hands kneading dough at dawn, wide shot of the empty bakery warming with oven light, macro of a loaf crust crackling, final shot of the baker flipping the sign to 'Open', gentle acoustic score across the cuts
Number the beats and keep one location as the through-line. The score carrying across cuts glues the sequence together.
Three-shot reveal: a climber's chalked hands gripping rock in close-up, a vertigo wide shot from above the cliff face, then a summit silhouette against sunset clouds, wind building through the cuts
Escalate the framing (close, wide, silhouette) while keeping the subject constant. Cut logic beats shot count.
Five-shot city morning montage: espresso machine steam, a rolling metal shutter opening, pigeons scattering across a plaza, a tram gliding past reflective windows, a newspaper landing on a doorstep, ambient city waking sounds
Montages allow one shot per image with no shared subject, but they need one shared time-of-day and palette.
Two-shot match cut: a spinning basketball on a fingertip in a dark gym, cut to a spinning globe in a classroom as sunlight sweeps across it, whistle echo bridging the cut
Match cuts are the most art-directable two-shot structure: name the shared motion explicitly.
Kling v3 generates native audio with lip-sync in 5 languages. Quote the spoken line, stage the delivery, and keep character details consistent across prompts.
A market vendor holds up a peach and says 'Picked this morning, try it', warm smile, morning light through canvas awnings, crowd murmur and crate sounds behind
Short single lines keep lip-sync tight. The staging ('warm smile') shapes the voice performance.
A ship captain at the wheel in rain gear squints at the horizon and mutters 'Storm's coming in early', spray hitting the windows, radio static, deep hull creaks
Muttered lines against textured ambience test audio nuance; keep the ambient layers to two or three named sounds.
A young chess prodigy slides her queen forward, looks up and says 'Checkmate in three', quiet tournament hall, clock ticking, a spectator's chair creaking
Confidence reads through pacing. A short declarative line after a physical action lands better than a monologue.
Vertical 9:16 clip: a florist wraps a bouquet and says 'These always sell out by noon', paper crinkling, scissors snipping stems, soft shop bell in the background
Mention 9:16 directly for vertical output, and keep the speaking subject centered for phone screens.
Kling handles stylized looks, from anime to painterly, with unusually stable character detail. Name the style tradition and keep it consistent across a project.
Anime style: a courier girl on a red scooter weaves through a rainy neon city, reflections smearing across wet asphalt, her scarf streaming, upbeat synth soundtrack
Name the style tradition ('anime style') first so every later detail is interpreted inside it.
Watercolor animation of a paper boat drifting down a rain gutter stream, pigment blooming at the edges of every shape, gentle piano, soft paper texture throughout
Medium-specific artifacts ('pigment blooming') convince more than the style name alone.
Stop-motion style: a tiny clay robot waters a windowsill garden, fingerprint textures visible in the clay, slightly jittery 12fps movement, cozy hummed melody
Asking for the medium's imperfections (fingerprints, 12fps jitter) is what sells stop-motion.
Ink wash animation of a crane taking flight over misty mountains, brushstroke trails following the wingbeats, negative space dominating the frame, single flute note held long
Styles built on negative space need you to protect it: say 'negative space dominating the frame'.
The most reliable Kling v3 prompt structure is subject and action, environment and lighting, camera, style and mood, then audio. Long 15-second takes reward one patient camera move over several rushed beats.
Kling v3 cuts between up to 6 shots in a single generation. Say 'four-shot sequence', give each beat one camera setup, and keep one subject, location, or sound as the through-line across the cuts.
Kling v3 generates lip-synced audio in 5 languages. Put spoken lines in quotation marks, keep them to a sentence, and stage the delivery (mutters, grins, calls out) so the performance matches the scene.
For stylized clips, lead with the tradition (anime, watercolor, stop-motion) and then ask for that medium's characteristic imperfections, like pigment bloom or clay fingerprints. The flaws are what make the style read as real.
For characters that recur across clips, reuse the exact same descriptive sentence in every prompt, and attach reference images in VideoGen when identity must hold precisely.
“VideoGen is awesome for scaling video production and improving turnaround time... The complex process of video editing, which could take days or months, now takes minutes!”


“VideoGen solves the biggest pain points of video production—complexity, cost, and time. In just a few clicks, anyone can create professional, copyright-free videos.”

“VideoGen is the most underrated tool for content creators who want to put out the highest quality content in the shortest amount of time.”

Sign in to generate clips from any prompt in this library, then build them into finished videos with narration, music, and captions.