A curated library of prompts for ByteDance's Seedance 2.0, organized by use case and shown with real generated outputs. Copy any prompt, adapt the details, and run it in VideoGen.
Expressive, physical motion is Seedance 2.0's signature strength. These prompts describe weight, fabric, and momentum so the model has real physics to animate.
Dynamic multi-shot chase: a parkour runner vaults across Moroccan rooftops at dusk, camera whips between wide drone shots and close handheld follows, footsteps and wind audible
Generated with Seedance 2.0 in VideoGen
Naming both camera distances ('wide drone shots and close handheld follows') tells Seedance where to cut inside one generation.
A surfer carves through the barrel of a turquoise wave in slow motion, spray glittering in backlight, camera tracking alongside inside the tube, roaring water fading to a muffled hush
Generated with Seedance 2.0 in VideoGen
Describing how the sound changes ('roaring water fading to a muffled hush') steers Seedance's native audio along with the picture.
A breakdancer spins into a one-handed freeze on a subway platform, jacket flaring with the rotation, fluorescent light flicker, sneakers squeaking on concrete, beat echoing off the tiles
Generated with Seedance 2.0 in VideoGen
Fabric and momentum cues ('jacket flaring with the rotation') give the model real physics to animate instead of a generic spin.
Vertical 9:16 clip: a trail runner splashes through a shallow creek in slow motion, water droplets frozen mid-air around her stride, morning backlight, rhythmic breathing and footfalls
Freezing one physical detail ('droplets frozen mid-air') anchors the slow motion. Mention 9:16 directly for vertical output.
A border collie sprints through a wheat field at golden hour, camera racing low alongside at ear height, stalks whipping past the lens, panting and wind in the audio
Stating the camera height ('racing low alongside at ear height') sells speed better than the word 'fast' ever does.
Seedance 2.0 generates stereo audio with lip-sync in the same pass as the footage. Put spoken lines in quotes and stage how they are delivered.
Warm kitchen scene: a grandmother teaches her grandson to fold dumplings, she laughs and says 'Not too much filling, patience!', flour dust catches the window light
Generated with Seedance 2.0 in VideoGen
Staging the emotion ('she laughs and says') shapes the lip-synced voice performance as much as the words do.
A head chef leans over the pass, tastes the sauce and calls out 'More lemon, thirty seconds!', clattering pans and ticket printer chatter, steam rising under heat lamps
Generated with Seedance 2.0 in VideoGen
Short imperative lines keep lip sync tight, and layered kitchen ambience makes the spoken line sit naturally in the stereo mix.
A deep sea submersible pilot leans toward the viewport and says 'There it is, the old wreck', beams of light cutting through dark water, sonar pings echoing softly
Generated with Seedance 2.0 in VideoGen
A quiet line over a low ambience bed tests audio nuance. Keep hushed dialogue to one sentence so the sync stays clean.
Vertical 9:16 clip: a street food vendor hands over a skewer and grins 'Careful, it's hot!', sizzling grill, night market chatter and lantern light
Small, specific reactions read as real. A grin and four words beat a paragraph of scripted enthusiasm.
Commercial-style prompts for product heroes and social ads. Swap the bracketed product for your own.
Luxury perfume commercial: liquid gold ribbons swirl around a crystal bottle in slow motion, then snap into the logo etching as orchestral swell peaks
Generated with Seedance 2.0 in VideoGen
Timing the visual snap to the orchestral peak works because Seedance generates picture and score in the same pass.
Vertical 9:16 clip: a neon running sneaker drops into a shallow puddle in extreme slow motion, crown splash rising around the sole, bass hit on impact, city lights bokeh behind
Generated with Seedance 2.0 in VideoGen
Syncing a bass hit to the moment of impact makes a generated clip feel edited rather than raw.
Macro shot of cold brew coffee poured over a sphere of ice in a crystal glass, slow swirls of cream blooming through the dark coffee, soft cafe jazz, morning window light
Generated with Seedance 2.0 in VideoGen
One hero action per clip. Cream blooming through coffee is a fluid-physics showcase Seedance handles cleanly.
Studio hero shot of [a titanium smartwatch] assembling itself from floating components in mid-air, each piece clicking into place on the beat, dark backdrop with a single light sweep
Assembly-on-the-beat prompts work when every component click lands on the music. Describe the rhythm, not just the parts.
Seedance 2.0 can cut between shots inside a single generation. Number the beats, give each shot one camera setup, and let the audio carry across the cuts.
Three-shot sequence: a fisherman pushes his boat off a misty shore at dawn, then a wide aerial of the boat crossing glassy water, then a close-up of his hands pulling in a net heavy with fish, gulls and lapping water throughout
Generated with Seedance 2.0 in VideoGen
Numbering the beats ('three-shot sequence') and giving each one a single camera setup is the most reliable multi-shot structure.
Multi-shot story: a paper lantern is lit in close-up, released into a night sky filled with hundreds more, then the camera rises to reveal the festival crowd below, warm strings swelling across the cuts
Generated with Seedance 2.0 in VideoGen
Audio that carries across the cuts ('swelling across the cuts') ties a multi-shot sequence together into one story.
Two-shot reveal: extreme close-up of a chess player's eyes narrowing, then a top-down shot as she slides the queen across the board and taps the clock, ticking clock and hushed tournament room
A two-shot with a clear match cut (eyes, then the move) is the simplest sequence to art-direct in a single prompt.
Multi-shot sequence: a snowball rolls downhill growing larger shot by shot, finally bursting against a pine tree in a slow-motion powder explosion, playful orchestral hits marking each cut
Escalation works when one subject stays the through-line. Let the scale change shot by shot, not the subject.
The most reliable Seedance 2.0 prompt structure is subject and action, environment and lighting, camera, style and mood, then audio. Be specific about the physical event you want to see; expressive motion is where the model shines.
Seedance 2.0 generates lip-synced stereo speech when spoken lines appear in quotation marks. Keep lines short, stage the delivery (laughed, whispered, called out), and describe the ambience around the voice so it sits naturally in the mix.
Seedance 2.0 handles multi-shot sequences in a single generation. Say 'three-shot sequence', give each beat one camera move, and keep one subject or sound as the through-line across the cuts.
Native audio is generated in the same pass as the footage, so name the soundscape explicitly: sizzling grill, wind gusts, a bass hit on impact. Describing how the sound changes over the clip steers the mix.
Seedance 2.0 can reference up to 9 images, 3 video clips, and 3 audio clips in one generation. When a text prompt is not enough to pin down a character or product, attach references in VideoGen instead of over-describing them in words.
âVideoGen 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!â


âVideoGen solves the biggest pain points of video productionâcomplexity, cost, and time. In just a few clicks, anyone can create professional, copyright-free videos.â

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

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