Seedance 2.5 on @higgsfield_ai Prompt: Generate a 30-second, 16:9, 24fps high-octane Eastern martial arts short film. High-budget cinematic live-action with AAA CG visual quality. No subtitles, no UI, no logos, no text, no voiceover, no background music—retain only realistic sound effects: roaring wind, sharp sandals skidding, silk whipping through air, katana blades slicing air, thunderous metal collisions, bamboo stalks snapping, heavy body impacts, whip pans, and fierce breathing. 【Visual Style & Characters】 Protagonist: A solitary female ronin in dark indigo-black traveling robes with crimson lining, hair tied back in a high messy ponytail, wearing silver forearm guards and carrying a dark steel katana. Cold, lethal, unyielding aura. Antagonists: Two distinct male dual-wielding shadow assassins in ash-gray and bone-white ninja cloaks with bamboo face masks. Gray Assassin A wields a heavy naginata (polearm) for long-range overhead sweeps and thrusts. Gray Assassin B wields twin kama (scythes) for rapid, ultra-close acrobatic slashes. Must remain two distinct individuals throughout. 【Combat Dynamics】 A relentless 2 vs 1 high-speed encirclement battle. Continuous hyper-speed ninja agility and swordplay from the very first frame—no stance-holding, no eye-contact pauses, no dramatic posing. Every 0.2–0.6 seconds features a dynamic transition in attack angle, elevation, or direction. Every jump, flip, slide, and rotation is seamlessly integrated directly into an attack, block, or parry. 【Camera Work】 Feral, intimate, hyper-dynamic handheld camera work dragged through the fight. Incorporates low worm's-eye views, dynamic side tracking, short rotating orbits, fast dolly-ins, violent recoil jerks on heavy weapon hits, Dutch angles (45°/85°), and foreground occlusions (snapped bamboo, flying leaves). Camera occasionally lags behind the action slightly before snapping back, conveying extreme momentum. 【Environment】 A misty, dense bamboo forest at twilight with falling autumn leaves and scattered mossy stone lanterns. The environment reacts heavily to the fight: bamboo stalks sheared in half, stone lanterns shattered by missed strikes, fallen leaves swirling violently in weapon draft currents, and dust trails kick-started by skidding feet. 【Action Timeline】 0–6s: Immediate dive into battle. Naginata lunges from foreground with a 0.1-second near-miss speed ramp skimming protagonist's cheek, while Scythe Assassin attacks from the rear flank. Protagonist low-slides beneath the naginata shaft, performs a one-handed ground spin to deflect the scythes, and parries the returning naginata thrust. 6–12s: Continuous high-low pressure. Protagonist vault-steps off Naginata's pole shaft into a mid-air twist slash, forcing Scythe Assassin back, lands into a cartwheel dodge, and side-flips through an attack gap with a low ankle-sweep cut. 12–18s: Midpoint chaos. Surrounded in a tight pinch. Camera fast dolly-in to an orbital whip-around as protagonist drops flat under a synchronized cross-strike, explosive body spin disarming one scythe, and back-slashing the polearm wielder. 18–24s: Maximum speed eruption. Multi-directional redirection. Head-tilt dodge → hilt-strike to pole shaft → 360° backflip counter → bamboo-stalk wall-rebound → aerial down-slice. 24–30s: Climax execution. Protagonist slices through the center gap. Rapid double-slash mauls Scythe Assassin, blasting him into a bamboo cluster that splinters on impact. Protagonist instantly pivots, executing a full-body rotating heavy cleave against Naginata Assassin across swirling leaves, closing for a final decisive strike. Final 0.5s: Protagonist stands alone, sword lowered, leaves settling around her, fast dolly-in to her cold eyes, hard cut to black. 【Negative Prompt / Exclusions】 Stance-holding, slow-motion pauses, turn-based fighting, idling, prolonged staring, floating energy beams, magic glowing auras, UI, text, logos, subtitles, character merging, cloned extra limbs, studio backdrops.

This downhill prompt is absolute chaos. 🚵‍♂️💀 Speed, rain, cliffs, broken terrain… and a final jump that should NOT work. Exactly the kind of high-energy sequence I love pushing with Seedance. Full prompt below 👇 A fearless downhill mountain biker charges through the final section of a world cup race on a forested cliffside during heavy rain, roots and rocks tearing at the tires as the rider jumps over a fallen tree, rides along a nearly vertical rock face and overtakes a rival by crossing through a narrow waterfall passage; the trail suddenly collapses ahead, revealing a massive ravine, forcing the rider to accelerate instead of brake, use a broken wooden bridge as a ramp and launch across the gap while the rear wheel separates midair; dense alpine forest, cliffs, spectators and storm clouds, kinetic extreme-sports , aggressive helmet-cam transitions, wheel-level tracking, aerial pursuit, sharp terrain parallax, realistic mud spray, cold grey daylight, green forest tones, white water and orange sparks, reckless survival intensity, ending with the bike landing on the front wheel alone, sliding sideways through the finish gate as the rider jumps free and the broken frame explodes behind them.

A Prompt for dynamic camera movement! Prompt ⤵️

Reverse calligraphy as a cinematic title sequence! Prompt ⤵️

Made with Seedance 2.5 + Gpt image 2 Prompt: Seedance 2.5-30-Second Ultra-Photorealistic Travel Vlog Prompt Create a 30-second ultra-photorealistic travel vlog that looks exactly like genuine footage spontaneously recorded by a real European travel creator on a high-quality handheld vlog camera. The same young European woman must remain visually identical throughout the entire video: identical face, facial proportions, hairstyle, hair color, skin tone, body proportions, age, wardrobe, accessories, and overall appearance. Maintain strict character consistency across every shot and camera angle. The woman should have natural skin pores, subtle facial imperfections, realistic eyes, individual hair strands, authentic hair movement, natural body proportions, realistic hands, and spontaneous human expressions. No beauty filter, no plastic skin, no excessive makeup, no perfect symmetry. The entire video must feel like real documentary/travel-vlog footage, not a commercial, music video, cinematic film, 3D render, or AI-generated sequence. CONTINUITY / CHARACTER LOCK Same woman in every shot. Young European woman, approximately mid-20s, naturally attractive but completely realistic. Natural complexion with visible pores and tiny imperfections, realistic eyes and eyelashes, natural eyebrows, slightly imperfect hair strands moving independently in the breeze. Same hairstyle and clothing throughout. Authentic posture and walking rhythm. No facial identity drift, no changes in age, hairstyle, clothing, accessories, body shape, or skin tone. She carries a small compact vlog camera, held naturally at arm's length during selfie shots. Her grip, fingers, hands, and camera position must remain anatomically realistic. LOCATION / ATMOSPHERE A beautiful European historic old-town district in the morning. Narrow stone-paved street, textured centuries-old buildings, colorful but realistic façades, wooden shutters, small cafés, outdoor tables, bicycles, bakery storefront, pedestrians, subtle street activity. Soft morning sunlight with realistic shadows. Natural atmospheric perspective. Authentic European street ambience. Background people behave independently and naturally, without staring at the camera. No duplicated pedestrians or repeated background characters. SHOT 1 — 0–5 SECONDS Selfie-style handheld vlog footage. The woman walks naturally down the old-town street while holding a small vlog camera at arm's length. She looks directly into the lens with a subtle, genuine smile. Camera has realistic handheld movement: slight walking bounce, tiny wrist movements, occasional micro-shake, natural framing imperfections, subtle breathing-related movement. Do NOT stabilize the footage perfectly. Her hair moves naturally as she walks through a light morning breeze. Her expression remains relaxed and spontaneous. She occasionally blinks naturally. Behind her: pedestrians walking casually, parked bicycles, café entrances, historic textured buildings and morning activity. Natural sunlight falls across her face as she moves between partial shade and sunlight. Preserve realistic exposure and skin tones. SHOT 2 — 5–10 SECONDS She naturally turns the camera away from herself toward the street. Create an imperfect handheld pan, as if a real travel creator casually redirected their vlog camera. Reveal: - detailed historic European architecture - stone façades and windows - outdoor café tables - people drinking coffee - bicycles moving through the street - a small neighborhood bakery - subtle morning street activity The pan should NOT be perfectly smooth. Include realistic wrist acceleration, tiny framing corrections and slight motion blur. As the camera moves from a darker section of the street into direct morning sunlight, produce a subtle automatic exposure adjustment. Highlights and shadows react naturally like a real camera sensor. No cinematic transition. No artificial camera glide. SHOT 3 — 10–15 SECONDS Cut naturally back to selfie footage. The same woman continues walking toward the café. She briefly says something casual and conversational to the camera, such as: “This little street is so beautiful.” Her lip movement must synchronize naturally with the words. Facial expressions are relaxed and understated. No exaggerated acting. She briefly glances toward something in the street and then back at the lens. She blinks naturally. A light breeze moves several individual strands of hair. Pedestrians pass naturally in the background without acknowledging the camera. Maintain realistic handheld shake, authentic walking bounce, natural breathing, realistic depth of field and believable street ambience. SHOT 4 — 15–21 SECONDS Transition naturally into POV footage from her camera. She approaches and enters a small independent local café. The camera moves through the doorway at normal human walking speed. Slight vertical bounce and subtle wrist movement remain visible. Inside: - warm natural café lighting - wooden tables - textured walls - realistic glass reflections - small background conversations - customers moving naturally - barista activity in the distance - subtle environmental shadows The camera approaches a wooden table containing a freshly prepared coffee and a small pastry. Show subtle realistic steam rising from the coffee. Reflections on the cup and nearby glass should obey real-world lighting. No exaggerated steam. No floating objects. No artificial-looking food. SHOT 5 — 21–26 SECONDS Cut back to selfie footage. The same woman is now sitting outside the café at a small table. She naturally picks up the coffee, takes a small casual sip, lowers the cup and looks into the camera. Her expression changes into a subtle authentic smile. Avoid influencer-style posing or exaggerated happiness. Maintain exact facial identity and wardrobe continuity. Background remains alive: pedestrians walking, bicycles passing, café customers talking, subtle movement of chairs and clothing in the breeze. Camera remains handheld and slightly imperfect, with realistic micro-shake. SHOT 6 — 26–30 SECONDS Continue the selfie shot for a natural ending. She keeps a relaxed smile for a moment, then her attention subtly shifts toward the street. The camera performs a small natural focus shift from her face toward the street behind her. Her face gradually becomes slightly softer while the background becomes clearer, like realistic autofocus behavior from a modern vlog camera. End spontaneously rather than dramatically. No cinematic fade, no artificial transition, no title card. CAMERA / PHYSICS Authentic handheld vlog-camera behavior throughout. Use: - realistic walking bounce - subtle wrist shake - natural micro-jitter - imperfect framing - physically accurate motion blur - realistic autofocus behavior - realistic exposure adaptation - natural depth of field - physically correct reflections - realistic shadows - natural skin response to sunlight - believable hair physics - realistic fabric movement - authentic pedestrian movement Camera should occasionally make tiny human corrections rather than moving perfectly smoothly. AUDIO / AMBIENCE Natural location sound only. Include subtle: - footsteps on stone pavement - distant conversations - café chatter - bicycle movement - light street noise - birds in the distance - door opening - subtle café interior ambience - faint cup movement - natural breathing Her spoken sentence should sound like a real casual vlog recording, with natural voice volume and slight environmental background noise. No dramatic soundtrack. No cinematic sound effects. ABSOLUTE REALISM REQUIREMENTS Prioritize photographic realism over visual perfection. The footage should contain the tiny imperfections normally found in genuine handheld travel footage. Avoid: - CGI appearance - AI-generated look - plastic skin - beauty filters - excessive skin smoothing - perfect facial symmetry - exaggerated expressions - overly cinematic lighting - perfect gimbal movement - impossible camera movement - warped hands - malformed fingers - distorted faces - inconsistent facial identity - changing clothing - changing hairstyle - duplicated pedestrians - frozen background people - unnatural walking - floating objects - impossible reflections - incorrect shadows - artificial depth-of-field - excessive bokeh - oversaturated colors - hyper-sharp artificial textures - dramatic transitions - commercial/ad-style posing Final target: The finished 30-second video must be indistinguishable from a genuine, spontaneously filmed European travel vlog captured by a real person on a premium handheld vlog camera. Every person, object, reflection, shadow, movement, facial expression, hair strand, exposure change and camera motion should obey real-world physics.