ByteDance via RunningHubGenerate 5–30 second videos from text or one opening image. Choose 480p, 720p, or native 1080p, with optional synchronized audio. Credits are quoted before generation.
Architectural Directives
Generate from a prompt or one opening image, direct longer clips with timestamped sequence beats, and choose optional synchronized audio across 480p, 720p, and native 1080p. Credits are quoted before generation.
Upload one strong still to anchor the subject, framing, color, and lighting, then describe only the motion and camera behavior you want to unfold.
Use the uploaded portrait as the opening frame. The subject turns slowly toward the window as warm light moves across her face; subtle handheld push-in, natural blink, restrained motion
Break a longer prompt into discrete second blocks ([0-8s], [8-16s], [16-24s], [24-30s]) to describe camera speed, character turning points, and lighting shifts in sequence.
[0-8s: Wide tracking shot through a foggy cedar forest] [8-16s: Push in as the subject notices a stone shrine] [16-24s: Slow orbit as blue spirit lights emerge] [24-30s: Low-angle tilt toward the canopy, then hold]
Enable generated audio when the video needs sound, then place concise cues such as footsteps, rain, or distant thunder beside the action they should accompany.
Heavy rain drumming against a corrugated warehouse roof, boots crunching on wet gravel, distant thunder at 12s, restrained cinematic reverberation, steady 24fps motion
30s Narrative Continuity Call Sheets
Built for longer directed shots and sequence pre-visualization. Start with text or one image, assign timestamped camera choreography, and choose a resolution that matches the draft or delivery budget.
Tell a complete brand story from raw ingredients to final packaged bottle in one continuous pass.
[0-10s: Macro shot of dew on hand-picked tea leaves] [10-20s: Camera glides toward a copper roasting kettle, steam rising] [20-30s: Settle on a matte black tea canister on polished slate, morning sun flare]
Use a clean opening portrait to guide identity and framing while adding restrained expression and camera motion.
Use the uploaded portrait as the first frame. Slow push-in, slight head turn, natural blink, loose hair moving in a soft breeze, stable facial proportions
Glide from an indoor studio through a doorway into a bustling exterior city plaza without warp artifacts.
[0-10s: Slow forward dolly tracking a fashion model through an empty brutalist concrete corridor] [10-20s: Model pushes open heavy bronze doors, stepping into blinding mid-day Mediterranean sun] [20-30s: Camera pulls up into a high crane shot revealing a crowded cobblestone market square]
Enable generated audio and describe the performance, ambience, and sound cues together in one directed shot.
Slow-motion performance of an acoustic cello player on a salt flat at sunset; camera orbits at a steady pace; synchronized bow scrape, resonant wood tone, and a soft coastal breeze
For longer takes, use timestamp markers (for example [0-8s], [8-16s], [16-24s], [24-30s]) to describe when character movement, camera direction, lighting changes, and sound cues should occur.
Establishes physical environment, lighting conditions, and character wardrobe before high-velocity action begins.
[0-8s]: Slow tracking shot across a foggy Victorian harbor at dusk. Gas lamps flicker against wet cobblestones. A female investigator in a dark wool trench coat steps off a moored steamboat, carrying a leather satchel. Soft foghorn in distance, ambient lapping water.
Transparent technical constraints and verified strengths to help you pick the right pipeline for each shot.
Describe an opening setup, mid-clip action, and intentional ending in a single request, with selectable durations from 5 to 30 seconds.
Upload a single opening frame to preserve its subject, framing, palette, and lighting while directing motion in the prompt.
Choose 480p for economical drafts, 720p for everyday output, or native 1080p when the final frame detail justifies the higher token cost.
Enable generated audio when the shot needs ambience or sound effects, and describe those cues alongside the corresponding visible action.
30-Second Continuity Pipeline
Start with text or one image, describe longer sequences with timestamped beats, then choose the duration, native resolution, and optional audio before generation.
Begin from text alone, or upload one image that establishes the subject and composition you want to animate.
For longer clips, divide the prompt into timestamp blocks that describe the setting, movement, camera pacing, and optional sound cues.
Choose the duration, aspect ratio, native resolution, and optional audio, then review the credit quote before submitting the job.
Keep exploring
Authoritative answers to generation parameters, resolution quotas, licensing, and prompt mechanics.
Explore the distinct creative directions behind notable AI video models.
2K clips, 5–15s, text or image to video
30-second stories, flexible references & precise editing
Action prompts, camera direction & motion references
Cinematic scenes, atmosphere & sound direction
Social concepts, style prompts & visual hooks
Materials, object interactions & visual experiments
Performance prompts, gesture & camera framing
