Action Vlog Heroine Transformation
A peaceful handheld street vlog escalates into a photorealistic superhero battle as its creator transforms into an armored heroine, defeats a giant shadow monster, and returns to address the camera.
Tag Collection
25 published prompts tagged “角色一致性”. Browse by scenario and copy in one click.
25 prompts
A peaceful handheld street vlog escalates into a photorealistic superhero battle as its creator transforms into an armored heroine, defeats a giant shadow monster, and returns to address the camera.
A man falls from a Tokyo skyscraper through a rapid sequence of Japanese dreamscapes before crashing back into Shibuya Crossing, with his face and proportions locked throughout.
A 30-second vertical fashion editorial uses four fixed outfit stickers, a mouse cursor, and full-body fly-in overlaps to transform one consistent protagonist through five K-pop-inspired looks.
A 30-second fantasy-comedy one-take follows an ancient forest spirit as he transforms into a man, joins a campsite, and delivers a mock-serious beer-commercial payoff.
A 15-second vertical food commercial keeps four family members consistent while coordinating macro product shots, Mandarin narration, character dialogue, and precise lip sync.
A high-energy urban fashion film follows a model through New York with an ultra-wide fisheye lens, extreme low angles, whip pans, jump cuts, light trails, and music-player overlays.
A handmade clay mug waits behind perfect porcelain cups in a 30-second stylized 3D short, using three image references, tightly locked character animation, and a 13-segment emotional storyboard.
The same 15-second one-take desert-rider prompt is tested across Seedance 2.5, Seedance 2.0, and MiniMax H3, with a fixed lateral tracking camera and evolving background action.
An unseen housemate washes dishes, cooks dinner, steals a fry, tucks its exhausted roommate in, and finishes the night with a cold beer in a cozy 30-second comedy short.
Two performers move through pink studios, blue spotlights, tinsel backdrops, glamour close-ups, and synchronized choreography in a tightly timecoded 30-second K-pop music video.
The complete fable of the proud hare and steady tortoise is compressed into a colorful 30-second family-oriented 3D animation.
A quiet European countryside day unfolds from sunrise to golden hour as a photorealistic observational documentary driven entirely by natural actions and environmental ASMR.
A single AI-generated fashion portrait becomes a 10-second Seedance 2.5 camera-movement shot without any text video prompt.
Three 30-second generations form one hand-drawn monster battle by recycling each segment's last frame as the next segment's starting image.
Two referenced executives confront each other on a rain-soaked skybridge in a tightly blocked corporate thriller with prop continuity and match-on-action cuts.
Two referenced characters move through a restrained countryside breakup scene with tightly tracked touch, jewelry continuity, blocking, dialogue, and ambient sound.
A Spanish podcast interview escalates from a hair-growth gag into a rugby-team brawl, then resolves with a shouted command and a feather transformation.
Five visual references anchor a 12-shot megacity pursuit with missiles, a barrel roll, magical sigils, consistent riders, and edit-friendly action beats.
A woman pulls the camera operator through six hard-edged portals in one unbroken handheld take, accelerating from a party to a silent mountaintop.
A 9:16 tracking shot transforms one woman through school, Lolita, streetwear, hanfu, futuristic, and stage looks using movement and environmental occlusion.
A complete three-minute action-film prompt choreographs a lone swordsman against more than twenty attackers with strict continuity, spatial logic, and sound design.
A single character reference anchors a 30-second cyberpunk black comedy built around hard cuts between noise and silence.
A character reference drives a comedic 3D commercial in which grapefruit juice transforms a desert into a summer sea.
A concise editing instruction removes every person except the protagonist while preserving the source shot.
Eight references define a continuous journey through six rooms with distinct emotions and visual styles.
They are reusable LLM prompt templates labeled with “角色一致性” in Easy Prompt, selected for practical workflows and clear structure.
Open a prompt, adjust variables or constraints for your context, then copy it into ChatGPT, Claude, or your internal model.
This page lists published prompts with the tag. Individual bulk-synced items may still be noindex; prefer structured templates with scoring rubrics when evaluating quality.