This media is not supported in your browser
VIEW IN TELEGRAM
Soon you’ll be able to:
• Lock a character with @-references (same face/outfit every shot)
• Attach sprite sheets & stills
• Drop in voice — upload or record
• Build multi-reference video that stays consistent
Images · Audio · References. One flow. ETA Not known at this point, but presumably in August?
Please open Telegram to view this post
VIEW IN TELEGRAM
👏1
Media is too big
VIEW IN TELEGRAM
- Reference(s): none
- Styles Support: Realism
"From above a boundless ocean of storm-lashed jungle canopy, the camera plunges through churning fog toward a tiny streak of light knifing between colossal tree trunks. Lock-on: a speeder skimming a raised root-highway at insane speed, rain hammering its hull in silver sheets. The camera slingshots ahead through a curtain of falling water, whips back, then drops tight to the rear thrusters — steam bursting off hot metal, shredded leaves and spray whipping past, warning lights strobing red through the mist. Wind shears a massive branch loose; it crashes down in an explosion of bark and orchids as the rider snaps a last-inch swerve beneath it, then threads through a cathedral of hanging vines, buttress roots, and a hollowed fallen trunk in one fluid line — startled macaws detonating outward in a blur of scarlet and blue. The camera darts through the same gaps, glued to the motion, wet leaves slapping the lens. One final bend around a moss-shrouded giant and sudden calm: the camera blasts outward past the treeline into a reveal of a vast river gorge, the storm breaking open above it, shafts of sunlight piercing the fog, and a thousand-foot waterfall feeding the river below — mist rising off the jungle floor and igniting into a rainbow."Please open Telegram to view this post
VIEW IN TELEGRAM
👍1
Please open Telegram to view this post
VIEW IN TELEGRAM
Please open Telegram to view this post
VIEW IN TELEGRAM
Media is too big
VIEW IN TELEGRAM
- Reference(s): none
- Styles Support: Realism
"[CONDITION DEFINITION]
15秒、1:1スクエア、共有用、参照画像なし。戦艦大和は全長約263メートル、全幅約39メートル、第二次世界大戦期最大級の超大型戦艦として描く。第1フレームから米海軍艦載機群による波状総攻撃の最中。多数の爆弾、魚雷、近接爆発による損傷が蓄積し、最後は船体中央部を起点とする一度の致命的大爆発で撃沈が決定する。明るくクリアな高品質実写戦争映画映像。鮮やかな青空、青から青緑の海、暗灰色の鋼鉄船体、白い水柱、橙色の爆炎、濃い黒煙を明確に色分離する。
[SUBJECT / SCALE]
主役は戦艦大和。長大で幅広い船体、巨大な主砲塔群、高い塔型艦橋、多数の対空砲を備えた実物大の超大型戦艦。艦載機は大和より明確に小さく、複数方向から次々に襲来する。大和は単独攻撃では沈まず、総攻撃の蓄積によって最後の致命傷へ至る。
[SHOT / FLOW]
0.0〜2.1秒、左前方からの高い斜め俯瞰ワイド。戦艦大和の長い全長を画面中央に大きく見せる。多数の小さな艦載機が左右と上空から高速で横切り、対空砲火、爆弾着弾、不規則な水柱が広範囲に発生する。カメラは艦載機を追わない。
2.1〜4.2秒、右側面寄りの中高度斜め俯瞰ワイド。艦首から艦尾まで読める角度を維持し、別方向の複数編隊が前景・中景・遠景を高速で交差する。甲板と船体周囲への攻撃が続き、火災と黒煙が増える。
4.2〜5.8秒、高高度の超俯瞰ワイド。カメラは固定。一機の小さな急降下爆撃機だけが画面上方から大和へ高速降下して接近し、爆弾を投下した後、側方へ一瞬で抜ける。大和の全長と艦載機の小ささを同じ画面で維持する。
5.8〜8.3秒、海面すれすれの超低空。カメラは波頭の直上を一直線に猛烈な速度で進み、海水の飛沫が画面手前とレンズ脇へ激しく跳ね上がる。前方の巨大な大和へ急接近し、魚雷本体は大きく映さず、海面上の白い魚雷航跡とともに左舷水線下へ迫る。命中と同時に船体より高い巨大な水柱が立ち上がる。
8.3〜12.2秒、左側面寄りの高い斜め俯瞰ワイドを一つの連続ショットとして維持する。大和を画面内の最大対象とし、前方、左右、上空から艦載機群が途切れず高速横断する。魚雷航跡、爆弾直撃、近接爆発、不規則な水柱、対空砲火が長い船体全体へ集中する。攻撃を止めず、火災、黒煙、損傷、速度低下だけを増加させる。12.2秒までは致命的大爆発を起こさない。
12.2〜15.0秒、同じ斜め側面ワイドと同じ戦闘の流れを維持する。攻撃が継続する中、蓄積した損傷が限界に達し、船体中央部への最後の直撃を起点として一度の致命的大爆発が発生する。独立した球形火球や複数段の爆発にはせず、船体中央から横長に噴き上がる橙白色の爆炎、濃い黒煙、衝撃波、水しぶきを一体化する。大和の船体と爆発の起点を同じ画面内に残し、カメラは大きく動かさず、少し広い全景で終了する。
[CAMERA / EDITING]
左前方斜め俯瞰 → 右側面寄り斜め俯瞰 → 固定超俯瞰の急降下攻撃 → 海面超低空의 魚雷追跡 → 左側面寄り斜め俯瞰の連続総攻撃 → 同一ショットの致命的大爆発。後方斜め俯瞰は使わない。艦載機を追跡せず、画面を一瞬で横切らせて速度を表現する。カメラが激しく直線移動するのは魚雷シーンだけ。
[SOUND]
効果音のみ。多数の航空機音、風切り音、対空砲火、爆弾炸裂、魚雷命中、水しぶき、鋼鉄のきしみ、連続着弾、最後の巨大爆発。BGM、台詞、ナレーションなし。
[NEGATIVE]
後方斜め俯瞰、艦載機を追い続けるカメラ、魚雷シーン以外の直線追従、寄りすぎた画角、艦載機が大和より大きく見える構図、似た俯瞰カットの反復、カット順の入れ替え、魚雷シーンの省略、12.2秒以前の致命的大爆発、総攻撃が途中で止まる展開、傾斜、転覆、船体破断、複数段の球形火球、船体から切り離された爆発、夜戦、暗くくすんだ色調、現代戦闘機、ジェット機、ミサイル、特攻、文字、字幕、ロゴ、人体損傷、死体、ミニチュア感、玩具感、安いCG表現。"Please open Telegram to view this post
VIEW IN TELEGRAM
❤1
PRO PROMPTS | SEEDANCE | GROK | WAN | HAILUO | HAPPYHORSE
Photo
Seedance 2.5 is now out in Dreamina (GLOBAL) - and im not going to lie to you guys its not cheap. SD 2.5 is 117% more expensive for same amount of seconds.
However if you choose to prompt in 480p then the per second cost is actually same.
- Right now there is only 480p and 720p available, no sign of 1080p or native 4k which were advertised
- The Beta test mode of 360 seconds is also not available yet it seems
I Will be releasing a showcase video later, showing some results and i will likely share a prompt here too.
However if you choose to prompt in 480p then the per second cost is actually same.
- Right now there is only 480p and 720p available, no sign of 1080p or native 4k which were advertised
- The Beta test mode of 360 seconds is also not available yet it seems
I Will be releasing a showcase video later, showing some results and i will likely share a prompt here too.
Dreamina
Dreamina: Free AI Video & Image Generator Powered by Multiple Models
Create AI video, image, and short drama with Dreamina for ads, social content, product visuals, and visual storytelling. Powered by leading AI models including Seedream, Seedance, GPT Image, Sora, and more.
Media is too big
VIEW IN TELEGRAM
- Reference(s): none
- Styles Support: Realism
"Style: Ultra photorealistic live-action, Hollywood blockbuster, IMAX, ARRI Alexa 65, Atlas Orion Anamorphic, HDR, premium Hollywood VFX, aggressive cinematic camera. References: image1: Keep the character identical. image2: Follow the storyboard framing and camera composition. Prompt: A 2.2m human-sized powered armor mecha crashes into an abandoned megacity, creating a massive impact crater. As the camera rises behind the hero, a colossal Titan Dragon towers above the skyline, crushing skyscrapers. The mecha raises its rail cannon as magnetic coils unfold, blue-white plasma rapidly compresses, lightning fills the sky, and nearby vehicles and steel structures levitate from intense electromagnetic force. The Titan Dragon charges a molten sun inside its jaws. After a brief silence, both unleash their ultimate attacks. A devastating blue-white rail cannon collides with a world-ending molten fire beam, creating a gigantic plasma sphere, nuclear-scale shockwaves, lightning storms, sonic booms, collapsing skyscrapers, debris tornadoes, and violent camera shake. The rail cannon pierces through the dragon, ending in a colossal blue-white explosion. The final aerial shot reveals the lone mecha standing motionless at the center of the destroyed city beneath crackling lightning and towering smoke clouds. Sound Design: Massive cinematic impacts, metallic landing, mechanical transformation, magnetic coil charge, electrical crackles, thunder, dragon roar, molten fire blast, colossal rail cannon, beam collision, plasma explosion, collapsing skyscrapers, trailer braams, epic choir, orchestral percussion, emotional brass finale. Negative Prompt: giant mecha, Gundam scale, redesign character, anime, cartoon, CGI, toy, weak VFX, static camera, stiff animation, low quality, text, subtitles, logo, watermark"Please open Telegram to view this post
VIEW IN TELEGRAM
Media is too big
VIEW IN TELEGRAM
- Reference(s): none
- Styles Support: Realism
"realistic footage sequence inside a moving subway car. The scene begins as grounded live-action realism, then gradually becomes physically impossible: water starts emerging and rising inside the train until the entire carriage feels like an ocean trapped indoors, while gravity subtly shifts and rotates, causing passengers, loose objects, and water movement to reorient in unnatural but visually coherent ways. The sequence must feel like premium cinematic footage with strong camera choreography, clear spatial continuity, and believable material behavior. [00:00–00:03] Open with a smooth forward tracking shot moving down the center aisle of a subway car in motion. Fluorescent lights flicker softly overhead, passengers sit and stand naturally, poles and seats create strong depth, and the windows show dark tunnel streaks outside. Keep the camera low and controlled at first, drifting past shoes, bags, and seat legs before rising into a medium aisle-level move. Establish the geometry of the carriage clearly. [00:03–00:06] Something begins to feel wrong. Water starts seeping upward from the floor seams and pooling unnaturally around shoes and bags. The camera transitions into a slow lateral drift along the aisle, revealing passengers noticing the change. Small objects begin to slide at slight angles as if gravity is no longer perfectly vertical. The camera gently tilts to echo the shift without becoming chaotic. [00:06–00:09] The event escalates rapidly. Water surges higher through the carriage, not as a simple flood, but as if an ocean is pressing into the train from within. The camera performs a careful three-quarter orbit around a central standing passenger while water rises around seats, poles, and hanging straps. Loose objects begin drifting sideways and upward as gravity rotates. Hair, coats, newspapers, and droplets all react differently depending on the changing pull. Keep the motion readable and spatially precise. [00:09–00:12] The train interior becomes an impossible submerged world. The camera glides deeper down the aisle in a long floating tracking move, now moving like it is suspended inside water while still preserving a cinematic live-action look. Some passengers are partially lifted against the walls, others brace against poles from new gravity directions, and streams of bubbles curve diagonally through the carriage. Water fills most of the space, rippling across the ceiling and windows as if the orientation of the whole environment is unstable. The camera should move with smooth, dreamlike confidence, emphasizing the impossible geometry. [00:12–00:15] Finish with the most ambitious camera move. The camera advances toward the far end of the carriage, then rolls slowly while pushing upward, revealing that the apparent floor, wall, and ceiling no longer behave normally. Water, passengers, and objects are all oriented along shifting gravity lines, yet the composition remains clear and elegant. End on a striking wide interior view of the subway car transformed into a sealed ocean chamber, with the entire space suspended between flood and rotating gravity, while tunnel lights streak past outside the windows. Realistic footage look, modern subway interior, believable wet materials, convincing water behavior, gradual transition from normal realism into impossible physics, strong lens-driven storytelling, premium camera movement, clean spatial continuity, no fantasy powers, no branded references, no copyrighted character references."Edit: Fixed the prompt, had wrong one
Please open Telegram to view this post
VIEW IN TELEGRAM
Media is too big
VIEW IN TELEGRAM
- Reference(s): Yes
- Styles Support: 3D CGI / Realism
"主角：(@image1) 一位身手敏捷的成年女性蜘蛛侠，体格健壮，身着贴身紧身衣，戴着或半戴着面具，行动迅速，身手矫健。她正被多名身着黑衣的忍者敌人追赶，在夜色中于屋顶间展开追逐。风格：电影级3D CGI，高端动作冒险画面，动态光照，精细材质，强烈的速度感，逼真的物理效果，流畅的动画，以及扣人心弦的屋顶氛围。夜幕下的城市天际线，霓虹灯的倒影，湿漉漉的屋顶，通风口、招牌、管道、电缆、屋檐、水塔以及楼宇间的狭窄缝隙。动作设计要清晰流畅，节奏紧凑。着重展现跑酷、闪避动作、空中转换、屋顶动量以及追逐压力。重要动作规则：女主角必须同时使用蛛网和绳索下降。蛛网用于快速移动、摆荡、改变动量以及短暂阻挡敌人。绳降是一种更可控的下降或跑墙动作，使用固定在上方或建筑物上的绳索。两者都必须在镜头序列中清晰出现。镜头：主要采用跟踪镜头，动作流畅有力但清晰易懂，偶尔使用一些独特的角度，例如俯视漂移、低空屋顶滑翔、侧向追逐和短暂的垂直向下跟踪。避免镜头过于复杂。优先考虑动作的清晰度和速度。0.0–4.0秒：女主角冲上一个已经移动的屋顶，全力奔跑，几名黑色忍者敌人出现在她身后。她翻越通风口，滑过管道，速度丝毫不减。跟踪镜头从前方拍摄，然后切换到侧视角，因为更多的忍者落在她身后。4.0–8.0秒：她到达屋顶边缘，发射蛛丝，荡过一条狭窄的街道缝隙。两名忍者紧随其后。她在最佳时机松开蛛丝，落地后翻滚并继续奔跑。镜头跟踪她的落地和向前恢复，显示追击者仍然紧随其后。 8.0–12.0秒 她向左横切过一个更高的平台，利用跑酷技巧在墙上滑行了几步，然后借助上方固定的绳索从建筑物侧面滑下。这次下降动作是可控且有策略的，并非荡绳。她沿着垂直的墙壁跑了一段，然后蹬地，落在较低的屋顶上，忍者们则慌忙追赶。使用向下追踪镜头清晰地展现垂直移动。12.0–16.0秒 在较低的屋顶上，女主角冲过一片布满管道、电缆和天线的杂乱区域。一个忍者几乎追上了她，于是她迅速向后发射蛛丝缠住他的腿，拖延他的时间。她没有停下脚步。镜头继续跟随，略微靠后并略微偏离中心。 16.0–20.0秒 她到达另一个平台，并采取了更激进的移动方式：先是短跳，一只脚踩在墙上，然后借助蛛丝快速荡过屋顶结构，改变动量。这为她的路线创造了一个流畅的弧线，扰乱了忍者的追击路线。一名忍者用力过猛，不得不重新调整。20.0–24.0秒 追逐再次向上推进。女主角爬上一个标志牌框架，翻过顶部，然后短暂地从更高的横梁上进行第二次绳降，穿过建筑物之间的缝隙。她用蛛丝控制下落，在半空中触碰墙壁，然后干净利落地落到狭窄的屋顶走道上。忍者们仍在追赶，一些在上方，一些在后方。24.0–27.0秒 她加速进入最后的逃脱阶段。忍者们从多个角度逼近。她将迅捷的跑酷动作与两次快速的蛛丝动作相结合：一次蛛丝让她更快地跨越缝隙，另一次蛛丝则短暂地阻挡了忍者的行进路线。整个动作应展现出一种绝望、迅捷和技巧娴熟的感觉。27.0-30.0秒，最终高潮。她从最后一个屋顶边缘跃下，发射出一根长长的蛛丝，荡入开阔的城市空间。在她划出一道弧线飞离时，一名幸存的忍者已经来不及赶到边缘。女主角消失在霓虹闪烁的城市天际线中，动作仍在继续，而追踪镜头则捕捉到她优雅的逃亡轨迹，直至最后一刻。音效：急促的呼吸声、脚步声、布料飘动声、屋顶碎片声、金属碰撞声、蛛丝发射的短促声响、可控的绳索摩擦声、忍者的脚步声以及强劲的电影追逐配乐。无对白。整个场景应始终聚焦于逃脱、追逐、速度和机动性。女主角从未真正停下来战斗。整个动作的核心在于通过巧妙的动作、跑酷、蛛丝穿梭和绳索速降来躲避追逐。"Please open Telegram to view this post
VIEW IN TELEGRAM
❤1🤣1😐1
https://www.youtube.com/watch?v=J7kwlGtgGb4
Seedance 2.5 Preliminary Review & Analysis
- This New version includes 30 second long videos, currently available in 480p & 720p. The 1080p and 4K expected to come later as the 360second beta test mode.
- The cost is 117% higher per second compared to Seedance 2.0 at 720p. If you prompt in 480p its basically same price as 2.0 although.
- It seems that 2D generations took nerf, while 3D generations are generally better. The distortion problem seems to be present on longer videos that go beyond 20 seconds.
- Storyboarding techniques seems useless now as SD 2.5 follows stuff differently. Based on few tests that I did.
Seedance 2.5 Preliminary Review & Analysis
- This New version includes 30 second long videos, currently available in 480p & 720p. The 1080p and 4K expected to come later as the 360second beta test mode.
- The cost is 117% higher per second compared to Seedance 2.0 at 720p. If you prompt in 480p its basically same price as 2.0 although.
- It seems that 2D generations took nerf, while 3D generations are generally better. The distortion problem seems to be present on longer videos that go beyond 20 seconds.
- Storyboarding techniques seems useless now as SD 2.5 follows stuff differently. Based on few tests that I did.
The Following platforms have it currently are Loova AI, Dreamina, JXP and CapCut
YouTube
Seedance 2.5 Review - The Ultimate Video Model or Flop?
The day is finally here after some delays. ByteDance has made a global launch of Seedance 2.5 to the masses, and we are going to test this new expensive model and see what the STRENGTHS and WEAKNESSES of this model are.
Get Seedance 2.5 Access now!
✨ Loova…
Get Seedance 2.5 Access now!
✨ Loova…
Media is too big
VIEW IN TELEGRAM
- Reference(s): none
- Styles Support: Realism
"During a stormy night, an ancient glowing door suddenly appears in the middle of a dense forest. A lone explorer approaches cautiously, holding a lantern. As they slowly open the door, a blinding golden light reveals a breathtaking world filled with floating islands, giant waterfalls, glowing trees, and a sky with two moons. The explorer takes one step through the doorway, and reality instantly transforms around them—ordinary clothes become mystical armor as magical energy swirls through the air. Suddenly, the ancient door vanishes forever. A distant voice whispers, “You can never return.” The explorer looks toward the endless magical horizon as the screen fades to black. Ultra-realistic visuals, cinematic fantasy, volumetric lighting, dramatic camera movement, seamless transitions, movie-quality VFX, emotional atmosphere, highly detailed environments, and an epic cinematic ending."Please open Telegram to view this post
VIEW IN TELEGRAM
❤1
This media is not supported in your browser
VIEW IN TELEGRAM
- Reference(s): none
- Styles Support: Anime
"0–4s: Rain pours over a futuristic neon city rooftop at night. Two cyber-enhanced anime warriors activate glowing energy blades as blue and magenta neon reflects off the wet surface.
4–8s: They sprint toward each other at impossible speed, exchanging dozens of sword strikes. Every clash releases sparks, holographic fragments, and glowing energy waves. The camera circles around them.
8–12s: Both teleport across rooftops, leaving digital afterimages and crackling electricity. Rain explodes into slow-motion droplets with every impact.
12–15s: They launch one final aerial attack. The collision creates a giant ring-shaped shockwave that sends neon signs, glass, and rain outward before fading into silence.
Style: High-end anime, cinematic lighting, realistic rain physics, neon reflections, smooth action choreography, premium VFX, no text, no watermark."Please open Telegram to view this post
VIEW IN TELEGRAM
❤1
- Native 4K
- Up to 30 Seconds of video
- 60B Parameter Architechture
- Identity Lock (Character consistency)
- Up to 6 shots in single creation (Not sure what this means exactly, but perhaps scene transitions)
- Multimodal support of Audio, Text, Images and Video
Stay tuned for a review on this.
Please open Telegram to view this post
VIEW IN TELEGRAM
This media is not supported in your browser
VIEW IN TELEGRAM
- Reference(s): none
- Styles Support: Realism / 3D CGI
"Cinematic fantasy epic, heavy volumetric haze + strong anamorphic halation, mystical waterfall cave atmosphere, ultra-pro film look, 35mm anamorphic lens, rich god-rays cutting through mist, high-end VFX, photoreal textures.
Use the provided image as the only visual reference. Lock exact faces, body types, hair, red textured catsuit and black robes of both characters. Environment remains the same mossy rock cave with cascading waterfalls and glowing lanterns.
0-2s: [Medium Close-up, slow push-in] Woman stands tall, suddenly extends both arms wide. A living film of glistening crimson scales rapidly crawls up her body from the feet, fully covering the red catsuit in overlapping reptilian armor. She aggressively flicks out a long forked serpentine tongue, eyes glowing with predatory intensity. Heavy haze + lens flare.
2-4s: [Dynamic Tracking, slight orbit] Scales finish sealing over her entire form with wet reflective sheen. She hisses, tongue still extended. Energy crackles across the new scaled skin.
4-6s: [Wide Shot cutting to Medium] Old man rises from the rock in one fluid motion and executes a sharp, powerful kata. Instant black smoke + violent whirlwind of dark energy erupts around him, completely engulfing his body. The vortex collapses and he emerges as a towering, muscular oni: deep crimson skin, massive curved horns, glowing yellow eyes, tattered black robes now fused with demonic armor, long white hair flowing wildly. Ground cracks under his feet.
6-7.5s: [Explosive Action Cut, low-angle tracking] Scaled woman lunges forward with blinding speed, claws extended. Oni meets her mid-air with a massive open-palm strike. Shockwave of red-and-black energy explodes outward, shattering nearby rocks and sending waterfall spray flying in slow-motion.
7.5-9s: [Handheld chaotic tracking] Brutal close-quarters exchange. She lashes with forked tongue and spinning scaled kicks that leave trails of red energy. Oni counters with sweeping horn strikes and black smoke tentacles that try to bind her. Sparks, debris and water explode with every impact.
9-10.5s: [Dramatic wide + rapid push-in] Oni leaps high, gathers a colossal black energy sphere above his head. Woman coils, then launches upward through the waterfall, body fully scaled and glowing. They collide in mid-air in a massive VFX climax — red scale shards and black smoke vortex collide, creating a blinding spherical shockwave that illuminates the entire cave.
10.5-12s: [Slow-motion closing action, orbiting camera] Final impact freezes for a beat in heavy haze and intense halation, both figures silhouetted against the exploding energy, water droplets suspended like glass, then the blast expands outward.
Photorealistic 8K, ultra-detailed wet scales and skin textures, perfect motion blur, coherent physics, heavy cinematic haze and anamorphic halation, high dynamic range, movie-level VFX, stable characters, no artifacts."Please open Telegram to view this post
VIEW IN TELEGRAM
❤1👍1
This media is not supported in your browser
VIEW IN TELEGRAM
- Reference(s): None
- Styles Support: Realism (tested)
"Audio Track Idea: Quiet ambient noise transitioning to Heavy Bass Drop plus Ticking Clock into Cinematic Hype Beat
0:00 - 0:02 | THE SETUP Visual: Real-life video. You stand still in a busy or open area like a street corner, balcony, or park, and begin throwing a very slow-motion punch or hand palm directly toward the camera.
Audio: Regular street noise, quiet and slow.
0:02 - 0:03 | THE TRIGGER Visual: Your fist hits full extension right in front of the lens. The motion briefly pauses or freezes for a fraction of a second as a small shockwave effect ripples out from your hand.
Audio: Sudden loud bass drop layered with a fast-ticking clock sound.
0:03 - 0:07 | THE SEEDANCE 2.5 TIME WARP Visual: The AI takes over the background. Time hyper-speeds into a fast time-lapse where the sun drops below the horizon, stars appear, clouds race across the sky, and city lights switch on fast. Meanwhile, you smoothly pull your fist back into a power pose in liquid slow-motion while realistic shadows and neon lighting shift dynamically across your face and clothes.
Audio: Massive, punchy beat drop with atmospheric synth sound effects.
PROMPT : Hyper-realistic cinematic shot. Keep main person's motion smooth in 60fps slow-motion pulling fist back from punch. Background transforms into a hyper-fast timelapse: sky shifts rapidly from sunset to night, clouds streak overhead, city lights flicker on rapidly. Dynamic moving light and shadows sweep realistic light across the person's face and clothes. Photorealistic, 8k resolution."Please open Telegram to view this post
VIEW IN TELEGRAM
Media is too big
VIEW IN TELEGRAM
- Reference(s): Yes
- Styles Support: 3D Realism (tested)
"Two adult female warriors (@image1 and @image2 engage in a fast, intense fistfight.
Warrior One is a tall, muscular woman with tied-back black hair, wearing fitted crimson-and-black sleeveless combat armor and leather hand wraps. Warrior Two is a lean, athletic woman with short silver hair, wearing a dark teal tactical bodysuit with reinforced gloves. Keep both women visually distinct and fully consistent throughout the video. They fight using only fists, blocks, parries, slips and evasive footwork. No weapons, no kicks, no powers.
They fight inside an abandoned circular stone arena at night. Rain falls lightly through a broken ceiling, forming wet reflective puddles on the floor. Damaged stone pillars surround the arena. Cold moonlight shines from above while warm firelight from torches around the edges creates dramatic contrast, rim lighting, smoke and reflections on the wet ground.
[00:00-00:03] Start with a low wide shot. Both warriors raise their fists and slowly circle each other, tense and focused. The camera dollies forward between them as rain falls.
[00:03-00:06] Warrior One explodes forward with a left jab, right cross and a powerful hook. Warrior Two slips under the hook, guards tightly and fires back with two compact body punches.
[00:06-00:09] Shoulder-level lateral tracking shot. Warrior One presses forward with a rapid barrage of straight punches. Warrior Two retreats across the wet floor, dodging and blocking while maintaining strong footwork and balance.
[00:09-00:12] Warrior Two parries a right hand outward and counters with a sharp hook to the ribs. Warrior One absorbs it, pulls in close, and both exchange a tight flurry of brutal short-range punches to torso and guard. Subtle camera shake on impact. Water droplets and sweat scatter naturally.
[00:12-00:14] Tight three-quarter angle. Warrior One throws a heavy overhand punch. Warrior Two pivots aside at the last second. Use a very brief slow-motion moment as the fist cuts past her face, with rain and sweat spraying through the air.
[00:14-00:17] Return to normal speed. The camera performs a controlled half-orbit as Warrior Two plants her feet and lands a clean rising uppercut. Warrior One stumbles backward through a puddle but stays standing. End on a medium-wide shot with both women separated again, fists raised, breathing heavily, rain falling between them, the duel unresolved.
Premium cinematic 3D CGI, realistic anatomy, polished action-game cinematic quality, believable body mechanics, grounded movement, clear readable choreography, realistic rain simulation, wet materials, dramatic lighting, subtle handheld energy, restrained motion blur, 16:9, no dialogue, no captions, no music.
Negative prompt:
extra arms, extra fingers, deformed hands, fused limbs, broken anatomy, duplicated characters, changing faces, changing outfits, kicks, weapons, magic, energy blasts, flying, unreadable choreography, excessive spinning, chaotic camera motion, heavy blur, text, subtitles, watermark"Please open Telegram to view this post
VIEW IN TELEGRAM
👍2
Flux 3 Video is now out https://x.com/bfl_ai/status/2084693191484469305
X (formerly Twitter)
Black Forest Labs (@bfl_ai) on X
FLUX 3 Video is here.
Serious, fun, creative, real, cinematic, whatever you need it to be.
Native audio, Text to Video, Image to Video with multiple frames, video continuation, dialogue in multi…
Serious, fun, creative, real, cinematic, whatever you need it to be.
Native audio, Text to Video, Image to Video with multiple frames, video continuation, dialogue in multi…
Media is too big
VIEW IN TELEGRAM
- Reference(s): None
- Styles Support: Realism (tested)
"STYLE: Photorealistic historical time-travel vlog, realistic ancient Maya ceremonial city, handheld smartphone selfie footage, massive stepped pyramid, limestone plaza, carved reliefs, torchlight, incense smoke, painted priests, dense crowd and pounding ritual drums. Dark comedy gradually becomes genuine panic. Show visible fresh blood, but no exposed organs, dismemberment or extreme gore. Maintain the girl’s face, clothing and phone consistently.
[00:00–00:04]
Handheld selfie shot. The modern girl walks through a crowded ceremonial plaza beneath a towering pyramid. Drums echo while flower petals and incense smoke fill the air.
GIRL:“Okay, I’ve been here five minutes, and apparently I’m already invited to a private ceremony.”
She turns the phone toward the cheering crowd.
[00:04–00:08]
Two ceremonial attendants guide her up the steep pyramid stairs. She smiles and waves awkwardly.
GIRL:“They gave me flowers, jewellery and a personal escort. I think I’m important.”
One attendant exchanges a serious look with the other.
[00:08–00:12]
At the summit, she turns the camera toward a large stone altar. Fresh dark-red blood stains its surface and slowly runs through carved channels onto the steps. A nearby ceremonial bowl contains blood.
Her smile disappears.
GIRL:“Okay… that is definitely fresh.”
A priest enters behind her holding a ceremonial obsidian blade with blood smeared along its edge.
[00:12–00:16]
Close handheld selfie. She quietly approaches a young bystander while keeping the bloody altar visible behind them.
GIRL:“Just checking—is this ceremony happening for me?”
BYSTANDER:“It is happening to you.”
A single drop of blood falls from the priest’s blade onto the stone.
[00:16–00:20]
Her eyes widen. She immediately runs down the pyramid, the phone shaking violently as attendants chase her. She slips slightly on a blood-marked step but keeps running.
GIRL: “Nope! Five stars for the architecture—zero stars for hospitality!”"Please open Telegram to view this post
VIEW IN TELEGRAM
❤1
The most realistic and production-ready video model. Generations at ZERO credit cost for 7 DAYS Up to 50 references, 30 seconds of perfect continuity, realism in physics, advanced VFX/SFX, and SOTA video-to-video editing 7-day unlimited access starts on August 7. Subscribe now and the unlimited models included with your plan are yours right away This one comes with a subscription only, it will not be in the Unlimited Models Marketplace
Get it here https://higgsfield.ai/pricing
I Thought i would be sharing this deal here. You won't get that many SD 2.5 unlimited deals, most of them are expensive,
There is of course something similar at Loova too. If you want to do short films or sreal length episodes, unlimited plans are must for anyone who doesnt have thousands of dollars to burn.
Please open Telegram to view this post
VIEW IN TELEGRAM
Media is too big
VIEW IN TELEGRAM
- Reference(s): Yes
- Styles Support: 3D Game Graphics
"Create a 25-second high-energy open-world crime-game sequence beginning directly from @image1 as the exact opening frame. Preserve the same playable Korean male character, red Pyongyang tram, socialist-era architecture, Korean signage, sunny daylight, realistic AAA graphics, dense pedestrians, tram wires, road layout, minimap and gameplay HUD. The sequence should feel like uncontrolled player gameplay rather than a cinematic trailer. Non-graphic video-game violence, exaggerated ragdoll physics, no blood or gore.
[00:00–00:03]
The character stands beside the red tram while the player quickly rotates the third-person camera around him, revealing the enormous central square, propaganda posters, pedestrians, traffic and distant monumental buildings. He begins jogging through the crowd, bumping shoulders with pedestrians as NPCs shout angrily in Korean.
[00:03–00:06]
He runs into the road, yanks open the door of a moving black sedan and throws the startled driver onto the pavement. The camera snaps behind the vehicle as he accelerates away, scraping along the side of the red tram and showering the street with sparks.
[00:06–00:11]
Fast chaotic driving through central Pyongyang. The sedan mounts the sidewalk, smashes through market stalls, benches and street barriers, then plows through a group of fleeing NPC pedestrians with exaggerated non-graphic ragdoll impacts. People dive aside, fruit crates explode across the road and police whistles erupt behind him.
[00:11–00:15]
The car launches from a wide stone staircase, flies over a propaganda parade rehearsal and lands sideways in the public square. It spins through banners and cardboard displays before crashing into a parked military truck. The windshield cracks, smoke rises from the hood and the character kicks the door open.
[00:15–00:20]
Several angry civilians and uniformed security officers surround him. A frantic street fight begins: quick punches, blocks, a spinning kick, one enemy thrown across the hood of the damaged sedan and another knocked into a stack of metal chairs. Nearby pedestrians record everything on their phones while others flee.
[00:20–00:23]
The character escapes through a narrow alley, vaults over food carts and climbs a fire-escape ladder. The camera follows tightly from below before swinging outward to reveal rooftops, apartment blocks, monuments, tram lines and the distant skyline.
[00:23–00:25]
From the rooftop, he jumps onto the roof of a passing red tram, barely catching the overhead structure. Police vehicles flood the intersection below as the tram speeds toward the giant central monument. End with the character standing triumphantly on the moving tram while the wanted level rapidly increases.
Realistic vehicle deformation, responsive NPC crowds, detailed environmental destruction, physical debris, convincing ragdoll motion, smooth third-person gameplay camera, sudden player-controlled camera turns, dynamic traffic reactions, Korean crowd dialogue, horns, sirens, tram bells, tire screeches, impacts and environmental city ambience. No cinematic black bars, no subtitles, no cuts away from gameplay."Please open Telegram to view this post
VIEW IN TELEGRAM
🔥3