پرامپت ویدیو با هوش مصنوعی؛ ۱۰ نمونه آماده قابل کپی و تست شده

برای ساخت ویدیوهای جذاب با هوش مصنوعی، فقط انتخاب ابزار مناسب کافی نیست؛ یک پرامپت ویدیو با هوش مصنوعی با جزئیات دقیق می‌تواند کیفیت نتیجه را تا حد زیادی افزایش دهد. در این مقاله، ۱۰ پرامپت ویدیو آماده را برایتان قرار داده‌ایم که می‌توانید مستقیماً کپی و استفاده کنید یا برای ایده‌پردازی و ساخت پرامپت شخصی از آن‌ها کمک بگیرید.

در ادامه با این موارد آشنا می‌شوید:

• ارائه ۱۰ پرامپت آماده و تست‌شده برای ساخت ویدیو (طبیعت، ترند، ورزشی و …)
• نکات کلیدی پرامپت‌نویسی برای گرفتن بهترین خروجی (نسبت تصویر، حرکت دوربین و …)
•  آموزش قدم‌به‌قدم اجرای پرامپت‌ها و ساخت ویدیو در هوشا

پرامپت ویدیو با هوش مصنوعی چیست و چطور یک پرامپت خوب بنویسیم؟

پرامپت ویدیو یک متن دستوری آماده برای توصیف و ساخت ویدیو است که در آن می‌توانید سوژه، صحنه، حرکت و اکشن، حرکت دوربین، نور و سبک تصویری را مشخص کنید. برای نوشتن یک پرامپت خوب، رعایت ۳ نکته زیر به شما کمک می‌کند:

1. توصیف دقیق سوژه: مشخص کنید چه مواردی در ویدیو دیده می‌شود و چه ویژگی‌هایی دارد.
2. تعیین سبک تصویری: سبک، فضای بصری و حال‌وهوای موردنظر را واضح بیان کنید.
3. مشخص کردن مدت‌زمان: مدت ویدیو را متناسب با خروجی موردنظر تعیین کنید.

همچنین می‌توانید پرامپت‌های آماده ساخت ویدیو را مطابق سلیقه خود شخصی‌سازی کنید. پلتفرم‌هایی مانند هوشا، انواع پرامپت‌های آماده برای ساخت ویدیو را در اختیار کاربران قرار می‌دهند.

۱۰ پرامپت آماده ویدیو با هوش مصنوعی

در ادامه، ۱۰ پرامپت تولید ویدیو با هوش مصنوعی آماده و قابل کپی قرار داده شده است. علاوه بر نتیجه هر پرامپت، تنظیمات مناسب و نام مدل هوش مصنوعی نیز برای استفاده و انتخاب راحت‌تر ارائه شده است.

۱. پرامپت ویدیو طبیعت و رودخانه

• نام مدل: Seedance 2
• نسبت کادر خروجی: 16:9
• مدت‌زمان: ۱۵ ثانیه

Photorealistic cinematic, one single continuous unbroken shot from start to finish — absolutely no cuts, no edits, no transitions, one fluid uninterrupted camera move, 16:9. Bright daylight over a wide river cutting through a lush forest, sunlight glittering and scattering across the rippling water, mist drifting off the rapids, tree-lined banks and overhanging branches softly blurred in the background. The shot begins low and direct behind a flashing silver fish darting and leaping fast through the shallows, the camera skimming just above the surface chasing close behind it as it weaves between sunlit ripples, rocks and breaking spray — erratic, lively and kinetic. Without any cut, in the same fluid motion, the camera keeps racing low over the water with the darting, leaping fish further down the river. Then, at the midpoint, a powerful white-headed eagle suddenly plunges in from above and the side, talons outstretched, and snatches the fish from the water — its claws piercing through and clamping around its body in one strike — and the camera sweeps with the strike in one continuous move. Still unbroken, the camera drives in onto the moment of capture and explodes into a dramatic bullet-time effect: time nearly freezes as the eagle’s talons clamp onto the fish in an extreme macro close-up, the fish’s body bending and arching in the talons’ grip, and the camera sweeps slowly around the frozen instant — shimmering droplets of river water bursting off the captured fish and hanging suspended motionless in mid-air, the fish’s scales and gills razor-sharp, the eagle’s talon texture, feathers and eye in crisp detail, the fish caught mid-thrash — hyper-detailed. One seamless continuous camera move — chase from behind, racing low over the river, into the eagle’s strike, ending in a bullet-time orbit around the catch. Flowing and dynamic, collapsing into near-frozen bullet time only at the macro catch. Shallow depth of field, strong motion blur on the chase resolving into crisp frozen detail, bright natural daylight, glittering water light, high dynamic range, ultra-detailed photorealistic textures — fish scales, water spray, feathers, talons — 4K, high-end wildlife documentary look. Pacing over 15 seconds: about 6–7 seconds of dynamic fish chase low over the river, the eagle striking around the midpoint, then the rest in slow bullet-time macro of the eagle’s talons gripping the fish. 15 seconds, single continuous take.

با تغییر رنگ بدن و سر عقاب، نورپردازی و زمان صحنه از روز به عصر یا شب، می‌توانید پرامپت را مطابق سلیقه خود ویرایش کنید. در صورت تغییر زمان، جزئیات مربوط به نورپردازی را نیز در سراسر پرامپت هماهنگ کنید. برای این ویرایش‌ها می‌توانید از چت جی پی تی کمک بگیرید.

۲. پرامپت ویدیو عاشقانه و کاپل

• نام مدل: Seedance 2.5
• نسبت کادر خروجی: 16:9
• مدت‌زمان: ۱۱ ثانیه

A young Iranian couple sits together on thick spring grass in a sunlit park clearing. A vintage transistor radio in the grass finds a soft Persian melody, she rests her head on his shoulder, and he kisses her forehead. Ten-second romantic sequence, no dialogue. ACTIVE REFERENCES @MAN: late-20s Iranian male, jacket open while seated. 100% matches the reference. @WOMAN: mid-20s Iranian female,100% matches the reference. LOCATION MAP Outdoor clearing in a quiet park, late afternoon. Ground is thick vibrant green grass. Behind the couple, midground and background: several trees with dark heavily textured trunks and branches, canopy full of bright light-green spring leaves. Small gaps in the canopy show pale blue sky. Low golden sun sits behind the couple and to camera-right, shining through the leaves toward the lens. The couple sits on the grass in the midground, roughly 3 meters in front of the nearest tree trunk.@MAN sits screen-left. @WOMAN sits screen-right, hip to hip with him, their shoulders touching.A vintage transistor radio rests upright in the grass within 0.5 meters of @MAN’s right hip, screen-left of him, its base pressing the grass blades flat. A small cluster of yellow wildflowers grows in the foreground grass immediately screen-right of @MAN’s hip, between the two of them and the camera. Camera stays on the same side of the couple for all four shots. Screen direction never flips: @MAN is always screen-left, @WOMAN always screen-right. SHOT A — 0:00 to 0:02 FIRST FRAME: the frame is already filled edge to edge with sunlit green leaves and dark branches. No empty sky frame, no blank opening, no title card. LENS LOCK SHOT A = 18°. 18° diagonal field of view, classic telephoto lens character, camera 6 to 8 meters from the branch. Strong background compression, distant leaves stacked flat behind the sharp foreground cluster, razor focus on one small group of leaves, everything else melts into creamy bokeh, close framing achieved through lens reach, not physical proximity. CAMERA: static tripod, lens angled slightly upward into the canopy, tiny slow drift no wider than 5 centimeters. Sun disc partly hidden behind a leaf at frame upper-right, breaking into soft golden flare and warm halation. ACTION: leaves sway in a soft breeze, individual leaves rotating on their stems with visible lag behind the gust. Sunlight breaks through moving gaps and sweeps warm hotspots across the leaf surfaces. In the lower third of frame, one heavily defocused warm bokeh mass suggests the seated couple far below; no faces, no readable bodies there. PHYSICS: wind moves branches first, leaves follow with delay and settle back; no synchronized robotic sway. LENS CHECK SHOT A: 18° maintained, no drift to wide-angle. HARD CUT at 0:02. SHOT B — 0:02 to 0:04 FIRST FRAME: the vintage transistor radio already fills the frame, sitting in the grass, tuning dial and worn metal speaker grille clearly visible. No empty grass frame. LENS LOCK SHOT B = 29°. 29° diagonal field of view, short telephoto detail lens character, camera 1.5 meters from the radio, close focus. Razor focus on the tuning dial, background grass dissolved into soft green and gold bokeh, radio pops sharply from the environment. CAMERA: static, lens at grass level, radio occupying the right two-thirds of frame, foreground grass blades crossing the lower edge slightly out of focus. ACTION: 0:02.0 to 0:02.4 the radio sits alone, dial steady, faint static hiss. 0:02.4 @MAN’s right hand enters from screen-left, white shirt cuff visible at the wrist. 0:02.6 to 0:03.4 his thumb and index finger grip the ribbed tuning wheel and rotate it slowly clockwise, the wheel turning with visible friction and small mechanical resistance. 0:03.4 the crackling static resolves into a soft Persian melody. 0:03.4 to 0:04.0 his fingers relax, ease off the dial and lift away screen-left.PHYSICS: the radio’s mass keeps the grass pressed flat beneath it and it does not shift or tip while the dial turns; finger pads deform slightly against the ribbed wheel; grass blades near the hand bend from the passing wrist and spring back. LENS CHECK SHOT B: 29° maintained, no macro distortion, no wide-angle coverage. HARD CUT at 0:04. SHOT C — 0:04 to 0:07 FIRST FRAME: both @MAN and @WOMAN are already fully visible in a medium close-up two-shot, seated hip to hip, shoulders touching, spatial relationship readable immediately. No empty frame, no delayed reveal. LENS LOCK SHOT C = 29°. 29° diagonal field of view, short telephoto portrait lens character, camera 4 to 6 meters from the couple. Both faces razor-sharp, tree trunks compressed closer behind them, canopy dissolved into creamy green-gold bokeh, the couple pops clearly from the environment. CAMERA: static on sticks, lens at their seated eye level, slight 3/4 frontal angle from camera-right. @MAN holds the left third of frame, @WOMAN the center-right third, warm bokeh negative space above them. ACTION: 0:04.0 to 0:04.6 both sit still, listening, chests rising with a slow shared breath. 0:04.6 to 0:05.2 @WOMAN’s mouth curves into a small closed-lip smile, her eyes softening and narrowing slightly, gaze drifting down and forward, not to camera. 0:05.2 to 0:06.2 she tilts her head slowly to screen-left and lowers it onto @MAN’s left shoulder, her temple settling against his suit shoulder. 0:06.2 to 0:07.0 she settles, one slow blink, and @MAN’s torso adjusts a few centimeters to take her weight. BODY AND GAZE: @MAN’s torso faces camera at 3/4, his head stays upright and forward through this shot. @WOMAN’s torso faces camera at 3/4, head tilts screen-left. Neither looks into the lens. PHYSICS: her long curly hair moves with delay behind the head tilt, mass swinging slightly past the stop before settling. Her knit sweater compresses where shoulder meets his wool jacket, and the jacket fabric folds under the contact. Grass under their weight stays flattened. LENS CHECK SHOT C: 29° maintained, no wide-angle expansion. HARD CUT at 0:07. SHOT D — 0:07 to 0:10 FIRST FRAME: both faces already fill the frame in close-up, her head already resting on his shoulder exactly as it ended in the previous shot. Same positions, same wardrobe, same hair state, no reset. LENS LOCK SHOT D = 18°. 18° diagonal field of view, classic telephoto lens character, camera 6 to 8 meters from the couple. Background compressed flat and dissolved into a soft warm color wash, razor focus on his eyes and her forehead, only the two faces are sharp, close framing achieved through lens reach, not physical proximity. CAMERA: static, lens at their eye level, faces filling the frame, his head screen-left and slightly higher, her head screen-right and lower against his shoulder. ACTION: 0:07.0 to 0:07.8 @MAN slowly tilts his head down toward screen-right, eyes lowering and locking on her face. 0:07.8 to 0:08.8 he holds the look, a faint smile pulling at his moustache, her eyes half-closed, both breathing slowly. 0:08.8 to 0:09.5 he leans down and presses his lips gently to her forehead at the hairline, contact held with light pressure. 0:09.5 to 0:09.7 he stays there, eyes closed. 0:09.7 to 0:10.0 the image fades softly to warm cream white.BODY AND GAZE: his eyes stay locked on her for the entire shot and never look at camera. Her gaze stays low and forward. Lips stay closed throughout, no speech, no mouth movement other than the kiss. PHYSICS: his neck and shoulder carry the lean with real weight transfer, her head shifts a few millimeters as he presses in and her hair compresses at the contact point. Loose curls at her crown drift in the breeze with soft delay. LENS CHECK SHOT D: 18° maintained, no wide-angle coverage, background stays compressed and dissolved. LIGHTING:Single natural key: low golden sun behind the couple and to camera-right, shining through the canopy. Camera stays on the shadow-facing side. Warm gold rim light traces the top of @MAN’s curls and the outer edge of @WOMAN’s hair, with visible halation bloom where the sun catches the hair edges. Faces are filled softly by warm bounce off the sunlit grass, staying gently exposed but never brightly keyed. Dappled leaf shadows move slowly across their shoulders and the grass. Lighting direction, sun position and shadow direction are identical in all four shots. No flat front key, no beauty fill, no studio light, no lighting flip after a cut.COLOR AND TEXTURE: Kodak Portra 400 film emulation. Warm nostalgic palette, deep saturated golden-yellows, creamy soft highlight roll-off with no clipped whites, shadows lifted with gentle warm brown undertones and never true black, saturation slightly pulled back on deep reds, cohesive faded analog wash. Fine real film grain, natural optical softness, no CG gloss, no game-engine look, no digital sharpening halos. AUDIO: Ambient park sound: soft breeze, rustling leaves, distant birds. Shot B: dry radio static that resolves at 0:03.4 into a soft Persian melody, santur and violin, thin warm mono transistor-speaker tone with slight lo-fi compression. The melody continues quietly under Shots C and D as diegetic radio sound only. No dialogue, no narration, no offscreen voices, no subtitles, no added orchestral score.

پرامپت ویدیو کاپلی برای حفظ استایل تصویر مرجع تنظیم شده است. برای شخصی‌سازی، کافی است تصاویر مرجع خود را بارگذاری کنید. در مورد موی سوژه زن نیز، در صورت نیاز، جزئیاتی مانند حالت و فر مو (Loose curls) را متناسب با سوژه تغییر دهید.  

پرامپت چیست؟

۳. پرامپت ویدیو عشق در طول زمان

• نام مدل: Seedance 2
• نسبت کادر خروجی: 16:9
• مدت‌زمان: ۷ ثانیه

SCENE CONTEXT Tehran railway station platform, 1943, daytime. A dense farewell crowd of soldiers and families fills the platform beside a departing steam train. A young soldier chases the accelerating train; a nurse leans from a carriage window; he catches the rear boarding steps just before a tunnel. ACTIVE REFERENCES @Image 1: [age] + [build/role] + [hair state: length, texture, how it’s worn] + [eye color] + [1–2 unique visible features]. 100% matches the reference. @Image 2: [age] + [build/role] + [hair state: length, texture, how it’s worn] + [facial hair or bare face] + [1–2 unique visible features]. 100% matches the reference. LOCATION MAP Foreground: platform edge, scattered luggage, well-wishers waving red-white-green lion-and-sun tricolor flags. Midground: the running man and the dark green and grey steel carriages of the Trans-Iranian Railway rolling left to right from the camera’s view. Background: station canopy ironwork, station lamps glowing warm gold, a massive black locomotive venting thick white steam, a dark tunnel mouth at the far right end of the platform. Steam haze visible at 15 meters depth, density 40%. Light comes from overcast daylight above, cut by warm practical glow spilling from carriage windows and station lamps. FIRST FRAME / BLOCKING @Image 1 centered in a medium frontal frame, mid-sprint toward camera, eyes locked forward, wide euphoric smile, crowd parting around him on both sides, flags waving in fore- and background, steam drifting through frame. The train rolls left to right beside him on the right side of frame. Everyone in frame is already moving. FORMAT MODE Timed multishot, 14 seconds total. Cuts only at the specified points, the camera does not cut on its own. OPTICS 0.0s–4.0s: MS, 47° neutral FOV, no drift mid-segment. 4.0s–7.0s: MCU, 29° portrait compression, no drift mid-segment. 7.0s–11.0s: alternating tight over-the-shoulder MCU at 29° and platform profile WS at 63°, no drift mid-segment. 11.0s–14.0s: WS, 63°, no drift mid-segment. Shallow depth of field throughout on the moving subjects; crowd and flags soften into bokeh in fore- and background. CAMERA Camera on the shadow side of the platform, operator axis parallel to the tracks. 0.0s–4.0s: steadicam pulling backward at 20 km/h at chest height, matching the runner, with subtle 1–2 cm handheld tremor. 4.0s–7.0s: smooth lateral tracking at 30 km/h in perfect tandem with the accelerating carriage, keeping the woman centered while station architecture streaks behind her. 7.0s–11.0s: fast pans and tracking at chest height beside the runner. 11.0s–14.0s: camera planted static at the platform edge, locked off, letting the train carry the subject away from the lens. Tonal look: wide latitude, gentle highlight roll-off, soft filmic color science. ACTION 0.0s to 4.0s — @Image 1 sprints at 25 km/h straight toward the retreating camera, weaving through embracing soldiers and well-wishers, shoulders dipping to dodge, arms pumping, boots striking the stone platform. Crowd waves tricolor lion-and-sun flags; steam rolls across the platform. 4.0s HARD CUT 4.0s to 7.0s — @Image 2 framed inside the heavy metal window of the moving carriage, leaning gracefully out, one arm extended back toward the platform. She looks back exactly once over her shoulder toward the platform behind the train, loose strands of hair whipping in the wind, then faces forward again, radiant and certain. 7.0s HARD CUT 7.0s to 11.0s — over the shoulder of @Image 1: the dark green and grey carriages gain speed to 35 km/h beside him. He pushes to a full sprint, jaw set, chest heaving, closing the gap meter by meter. INSERT CUT to a platform profile WS: his full figure streaking past blurred bystanders and flags. 11.0s HARD CUT 11.0s to 14.0s — static WS at the platform edge framing the rear boarding steps of the final carriage. @Image 1 reaches the steps, grabs the vertical exterior handrails with both hands, plants one boot on the lowest step and leaps, hauling his body weight up onto the moving stairs. The train carries him away from the lens toward the dark tunnel mouth, steam and steel swallowing the frame as the carriage enters the tunnel. PERFORMANCE @Image 1: exhilarated and breathless — wide euphoric smile in the opening, tightening into focused intensity in the chase: brow lowered, nostrils flared, visible sharp exhales fogging faintly in the cool air, sweat sheen at the temples, living eyes with bright catch-lights locked on the train. In the leap, muscles strain visibly through the uniform sleeves. @Image 2: assured joy — soft smile, chin lifted, the single backward glance held for one full beat with steady eyes, absolute certainty in the set of her shoulders. Pore-level skin realism on both faces. PHYSICS Full body weight and inertia in the sprint: uniform fabric bounces with each stride, brass buttons catch light on impact steps. The train has locomotive mass — slow build of momentum, steel wheels grinding on rails, contact shadows under the carriages. Steam behaves as heavy vapor, curling and shearing in the train’s slipstream. Flags snap in the wind. On the leap, the man’s body swings with real momentum as the handrails take his weight, boots landing solidly on the metal step. LIGHTING Overcast daylight from above at 5600K diffused through heavy industrial steam, soft wrapping key on faces. Warm 3200K practical glow from carriage interiors and station lamps rims the woman’s kerchief and spills across the platform. Steam catches the light in bright volumetric shafts between carriages. Exposure protects highlights in the white steam; faces sit half a stop bright against the misty platform. COLOR GRADE Nostalgic film emulation: deep teal held in the shadowed ironwork, steam and wet stone; rich warm amber in skin tones and the lamp spill; the red, white and green of the lion-and-sun flags reading as saturated fabric catching daylight against the desaturated smoky platform. Gentle halation on the warm practicals. WARDROBE @Image 1: khaki wool 1943 military uniform, rigid epaulets, high collar, polished brass buttons, fabric creased at the elbows from the sprint. @Image 2: structured grey-blue nurse dress, crisp starched white apron, white kerchief tied neatly over her hair — all fabric clean and pressed. AUDIO Heavy rhythmic chugging of the steam locomotive building in tempo, two shrill whistle blasts, the chaotic hum and cheering of the crowd, boots on stone, and a sweeping urgent orchestral score driving forward and swelling into the tunnel. STYLE Photoreal 1940s period cinema, fine organic film grain, soft halation, anamorphic character with gentle oval bokeh. OUTPUT SETTINGS 8K, anamorphic widescreen, real-time speed in every segment, 14 seconds total. POSITIVE LOCKS @Image 1’s face, hair and stubble stay identical to the reference in every cut; his face stays bare with eyes fully uncovered throughout. @Image 2’s face, kerchief and dress stay identical to the reference. The train moves left to right from the camera in every shot. Same overcast daylight, steam density and platform throughout. Tricolor lion-and-sun flags stay visible in crowd shots. Cuts only at 4.0s, 7.0s and 11.0s.

علاوه بر پرامپت بالا، ۱۴ نمونه ویدیوی دیگر با مضمون عشق دردوره‌های تاریخی متفاوت، در بخش استایل‌های آماده هوشا قرار دارد (ویدیوهای مشخص شده در تصویر پایین). این ویدیوها نمونه‌هایی از سناریوی «عشق در گذر زمان» هستند که شکل‌گیری و ادامه یک رابطه عاطفی را در دوره‌های زمانی مختلف به تصویر می‌کشند. در این نمونه‌ها، فضای بصری، پوشش شخصیت‌ها و محیط متناسب با دوره‌های مختلف (در ایران و جهان) تغییر می‌کند و می‌توان روایت‌هایی با حال‌وهوای ایرانی یا خارجی ایجاد کرد. برای ساخت یک ویدیوی کامل شما می‌توانید هر دوره زمانی را به‌صورت یک کلیپ مجزا تولید کنید و سپس کلیپ‌ها را به ترتیب زمانی به یکدیگر متصل کنید تا روند آشنایی و عشق دو شخصیت در طول زمان به‌شکل یک روایت پیوسته نمایش داده شود.

۴. پرامپت ویدیو ترند اینستا

• نام مدل: Seedance 2.5
• نسبت کادر خروجی: 9:16
• مدت‌زمان: ۱۵ ثانیه

Use @Image 1 as the exact model reference. Preserve her facial features and identity consistently throughout the entire video. Use @Image 2, @Image 3, and @Image 4 as three flat-lay outfit references. Dress the model in each outfit exactly as shown, matching the garments, colors, shoes, and accessories. Create a photorealistic, cinematic fashion outfit-change reel. Vertical 9:16 format. Duration: 17 seconds. CAMERA AND SETTING: One continuous overhead shot, with the camera mounted on the ceiling looking straight down, close to a ceiling fan. Keep the camera position and framing fixed throughout. The model lies on a classic Persian rug with a cream field, blue and red floral medallion patterns, and fringed ends. The rug rests on dark herringbone parquet flooring. A burgundy velvet sofa is partially visible at the frame edge. Use a warm cinematic color grade, moody golden tones, and a soft diagonal shaft of light from the upper left. Keep the rug, parquet, sofa, lighting, and framing consistent throughout. The model fills most of the vertical frame from head to feet, with her head toward the top and her feet toward the bottom. STARTING OUTFIT — IMPORTANT: From the very first frame, the model wears a fitted black short-sleeved turtleneck top, full-length tailored black trousers, and black closed-toe shoes. The trousers reach the shoes, fully covering her legs and ankles. Her feet are fully covered by the shoes. Do not start with a mini dress, skirt, shorts, bare legs, or bare feet. This fully covered black outfit is the fixed starting look before the three reference outfits appear. CEILING FAN — ONE SINGLE CONSISTENT FAN: Use one ceiling fan throughout. Its round center hub sits exactly at the LEFT EDGE of the frame, with half of the circular hub visible and the other half cropped outside the frame. The fan is a normal physical size, but its proximity to the ceiling-mounted camera makes the blades appear large. Each blade is long enough to sweep across the entire frame. After the model switches it on, the fan rotates smoothly CLOCKWISE, as seen by the camera, at one constant speed for the remainder of the video. It never reverses, slows down, pauses, or changes direction during any transition. All blades belong to the same fan and rotate around the same fixed hub. Never show a second fan, a second hub, or independently moving blades entering from opposite sides. OUTFIT-CHANGE EFFECT: Each outfit change is a gradual blade-driven wipe, not a sudden cut or an instantaneous full-body transformation. As a long fan blade sweeps clockwise across the frame, the area its leading edge has already passed reveals the NEW outfit. The area the leading edge has not yet reached still shows the PREVIOUS outfit. The blade’s leading edge is the moving boundary between the two outfits. The transformation progresses continuously across her entire body, including her trousers or skirt, accessories, and shoes, until the blade has passed her feet. Only her outfit changes. Her identity and the surrounding room remain consistent. SHOT 1 — 0–4 SECONDS: The model lies on the rug wearing the fixed starting outfit: a fitted black short-sleeved turtleneck top, full-length black trousers, and black closed-toe shoes. Her legs, ankles, and feet are covered from the opening frame. A remote lies on the rug a short distance from her hand. She rolls slightly toward it, picks it up, points it at the fan, and switches it on. Once the fan is rotating, she looks back toward the camera and smiles confidently. TRANSITION 1 — APPROXIMATELY 4–5 SECONDS: A blade sweeps clockwise from the half-visible hub at the left edge, gradually wiping the starting outfit into the complete outfit from @Image 2. The change follows the blade’s leading edge all the way to her feet. SHOT 2 — 5–8 SECONDS: The same model lies in the same overhead composition, now wearing the exact outfit from @Image 2. She holds an elegant, confident pose. TRANSITION 2 — APPROXIMATELY 8–9 SECONDS: A blade from the same fan continues rotating CLOCKWISE at the same constant speed. Its leading edge gradually wipes the outfit from @Image 2 into the exact outfit from @Image 3. SHOT 3 — 9–12 SECONDS: The same model remains in the same position and framing, wearing the exact outfit from @Image 3 and holding a confident pose. TRANSITION 3 — APPROXIMATELY 12–13 SECONDS: A blade from the same fan continues rotating CLOCKWISE at the same constant speed, gradually wiping the outfit from @Image 3 into the exact outfit from @Image 4. The wipe follows the blade’s leading edge across her entire body and finishes at her feet. SHOT 4 — 13–17 SECONDS: The same model remains in the same overhead composition, now wearing the exact outfit from @Image 4. After settling into her pose, she picks up her phone and checks it. The phone screen contains no readable text or visible branding. CRITICAL CONTINUITY: Preserve the model’s face and identity throughout. Keep one fan, one fixed half-visible hub at the left edge, and smooth constant clockwise rotation after activation. Never reverse the fan during the second or third transition. Keep the Persian rug, parquet flooring, burgundy sofa, diagonal light beam, and overhead framing consistent. No visible logos, captions, watermarks, or text anywhere in the frame.

این پرامپت با استفاده از ۵ عکس مرجع، شامل تصویر پرتره سوژه و تصاویر استایل‌های مختلف، ساخته شده است. بهتر است تصاویر استایل را برای حفظ تناسب، با ابعاد یکسان در ابزار ساخت تصویر بارگذاری کنید. برای تغییر جزئیاتی مانند فرش یا رنگ کاناپه نیز کافی است عبارت‌هایی مانند «Persian rug» و موارد مشابه را با توصیف دقیق فضای دلخواهتان جایگزین کنید.

آموزش پرامپت نویسی عکس از مقدماتی تا پیشرفته

۵. پرامپت ویدیو رقص

• نام مدل: Kling Motion
• نسبت کادر خروجی: 9:16
• مدت‌زمان: ۱۲ ثانیه

۶. پرامپت ویدیو ورزشی و استادیوم

• نام مدل: Seedance 2
• نسبت کادر خروجی: 16:9
• مدت‌زمان: ۱۵ ثانیه

Scene: One continuous drone shot of 2 men fighting Capoeira on a mountain top platform overlooking the Rio Favelas. Create a cinematic drone-shot sequence of two men performing an intense capoeira exchange on a circular concrete fighting platform perched on the edge of a cliff high above the Brazilian favelas. The platform is weathered concrete with chipped yellow paint markings. Massive mountain cliffs, dense colorful shanty buildings, and layered Rio-style mountain scenery surround the location. The mood is tense, athletic, and beautiful. Keep the motion fluid and continuous like a skilled drone operator is capturing the action in real time. Use realistic movement, natural body mechanics, and strong environmental scale. Style: Use image_3 as the style and aesthetic of the scene for the cinematorgraphy. Camera Movement: Use image_2 and follow the camera motion exactly as directed in the storyboard. Men Fight Sequence: Use Image_1 as the diagram to follow the sequence to create choreography of the fight scene. Start from exact view from image_1. We start high above the brazilian favelas looking down at the men in the ready position about to fight. Start with a very high diagonal aerial establishing shot looking down on the circular platform and both men squared up in ginga, tiny within the massive landscape. The drone descends in a smooth arc toward them, maintaining the overhead angle as the environment establishes danger, height, and scale. As the drone gets closer, reveal the men clearly in ginga, circling lightly and rhythmically. Continue pushing in. The lighter fighter begins a feint and entry while the darker fighter stays reactive and ready. The drone sweeps lower and closer to the platform edge, then glides into a low sweeping pass near the fighters as the exchange intensifies. Begin an orbit around them as the darker fighter drops into esquiva, slipping under the incoming motion. The orbit should continue leftward around the pair as the lighter fighter flows into meia-lua de compasso, carving a spinning heel kick through space. Keep the drone moving with parallax so the background mountains and favelas shift behind them. Tighten the orbit for a more intimate view as the darker fighter answers with a cocorinha counter, dropping low under the kick. Then continue the close orbit into the armada exchange: one fighter spins high while the other defends and recoils with awareness. The drone should stay dynamic, circling them with changing depth, sometimes one fighter dominating the foreground while the other stays in midground. Then transition into the au / flow transition. One fighter inverts through a fluid cartwheel while the other ducks and tracks the movement. Let the drone change depth and angle mid-orbit, drifting past one fighter’s shoulder and then widening slightly so the full acrobatic motion reads clearly. Finish by easing back into a wider orbit as both fighters reset and return to ginga, facing each other again in a controlled, rhythmic stance. End with the drone pulling slightly wider to reconnect the men to the circular platform, the cliff edge, the favela below, and the surrounding mountains. Overall style: cinematic realism, controlled drone cinematography, fluid capoeira motion, dynamic parallax, strong sense of height and danger, elegant fight choreography, natural lighting, subtle filmic texture, high-end action documentary feel.

برای استفاده از پرامپت ویدیو استادیوم در فضای آزاد، ابتدا یک استوری‌بورد ۸ مرحله‌ای به‌عنوان تصویر مرجع برای نمایش زاویه‌های مختلف ۲ ورزشکار بسازید. پیش از اجرای پرامپت، استوری‌بورد را با هوش مصنوعی عکس ساز در هوشا ایجاد کنید و سپس آن را همراه با پرامپت در ابزار ساخت ویدیو آپلود کنید.

۷. پرامپت ویدیو جام جهانی و فوتبال

• نام مدل: Kling 3.0 Turbo
• نسبت کادر خروجی: 16:9
• مدت‌زمان: ۱۵ ثانیه

STYLE: Polished animated comedy short, ~15 seconds, 16:9. Warm, soft studio lighting against a plain dusty-mauve wall. Wide/establishing shots in a soft 3D “Pixar-lite” render; reaction close-ups in clean 2D cel-shaded anime style. Exaggerated cartoon expressions, smooth lip-sync, light comedic timing. CHARACTERS (caricatures, comedic parody): – RONALDO: cartoon Portuguese footballer, slicked dark hair, thick brows,   smug grin, red-and-green Portugal #7 jersey, lounging with a tablet. – MESSI: cartoon Argentine footballer, spiky brown hair, light beard,   white-and-light-blue Argentina #10 jersey, more animated and earnest. SETTING: Both sit on a dark red/brown couch, a low wooden coffee table in front with stacked papers, a coffee cup, and a soccer ball on the floor. SHOTS & DIALOGUE (burn in English subtitles, white text, bottom-center): 1. Wide 2-shot. Messi turns excitedly to Ronaldo, who scrolls a tablet    looking bored. MESSI: “Hey, Cristiano!” 2. Same shot. MESSI (proud): “I scored 2 goals today! Not bad, right?” 3. Cut to close-up of Ronaldo, smug half-smile.    RONALDO (mock-congratulatory): “Only 2! Congratulations, amigo!” 4. Messi close-up, deadpan, arms crossed: “So… who won today’s competition?” 5. Insert: tablet screen showing a sports stats page (Ronaldo’s photo,    “GOALS: 2”). Cut to Ronaldo turning the tablet toward Messi, grinning.    RONALDO: “I scored 2 goals too!” 6. Messi close-up: shocked open-mouth beat, then breaking into a laugh.    MESSI: “Seriously? You couldn’t let me enjoy it for one day?” 7. Wide 2-shot, both laughing and fist-bumping.    RONALDO: “Enjoy it? I was just making sure we’re still tied hahaha!” AUDIO: Two friendly male voices with light accents (Portuguese / Argentine flavor), playful banter tone. Upbeat, lighthearted; both end laughing warmly.

با تغییر دیالوگ‌های بین دو سوژه، می‌توانید این پرامپت را برای تم طنز شخصی‌سازی کنید. برای تغییر کاراکترها نیز کافی است ویژگی‌های ظاهری و نام سوژه‌های پرامپت را با بازیکنان دلخواهتان جایگزین کنید.

۸. پرامپت ویدیو کودک

• نام مدل: Veo 3.1
• نسبت کادر خروجی: 9:16
• مدت‌زمان: ۸ ثانیه

{   “shot”: {     “type”: “single_continuous_shot”,     “camera_position”: “front_low_angle”,     “camera_movement”: “static_with_minor_handheld_vibration”,     “framing”: “child_full_in_frame_sitting_on_floor”,     “lighting”: “soft_natural_daylight”   },   “environment”: {     “location”: “indoor_living_room”,     “details”: “soft_colors,toys_around,cozy_background”,     “reporter_visibility”: “only_hand_and_microphone_visible_no_face”   },   “characters”: {     “child”: {       “age”: “1_year_old”,       “appearance”: “cute_happy_baby_sitting_with_small_movements”,       “expression”: “smiling_laughing”     },     “reporter”: {       “appearance”: “only_hand_with_microphone_enters_from_right_side”,       “voice”: “adult_iranian_male_or_female_natural”     }   },   “audio”: {     “ambient”: “soft_room_tone”,     “effects”: “baby_laughter_no_music”   },   “dialogue”: [     {       “time”: “00:02”,       “speaker”: “reporter”,       “text”: “تو کدوم بخش ارتباط با فامیل مهارت داری؟”     },     {       “time”: “00:05”,       “speaker”: “child”,       “text”: “تو بخش قطع ارتباط!”     },     {       “time”: “00:07”,       “speaker”: “reporter”,       “reaction”: “laugh”     }   ],   “style”: {     “tone”: “funny_warm_home_video”,     “realism”: “high”,     “duration”: “8_seconds”,     “aspect_ratio”: “9:16”   } }

با تغییر دیالوگ‌ها، افزودن توضیحات مربوط به لباس کودک، بارگذاری عکس مرجع و تغییر سن کودک، می‌توانید پرامپت را متناسب با سوژه و هدف خود شخصی‌سازی کنید.

هوش مصنوعی Veo 3.1 گوگل چیست؟

۹. پرامپت ویدیو ترند

• نام مدل: Seedance 2.5
• نسبت کادر خروجی: 16:9
• مدت‌زمان: ۲۰ ثانیه

SCENE CONTEXT A high-energy 20-second GTA VI gameplay teaser set in Vice City at night. This is raw in-game third-person gameplay footage from Rockstar Games’ GTA VI — not cinematic film. A controlled multi-shot sequence of disconnected gameplay moments joined by HARD CUTS: driving, shooting, walking the neon strip, police escape. No story, no continuity between segments — pure energetic teaser rhythm. ACTIVE REFERENCE The single protagonist in every segment is the character from the uploaded reference image. Face, hairstyle, skin tone, body proportions and outfit match the reference image 100% in all segments and all camera angles. The reference image is the absolute source of truth for identity — no drift between cuts, no alternate versions, no extra protagonists. RENDER STYLE LOCK The entire video looks like real-time gameplay captured from a modern AAA game engine at maximum graphics settings — GTA VI trailer-grade fidelity. Highly detailed but unmistakably a video game: game-engine skin shading, game-engine hair, crisp full-frame focus with no cinematic shallow depth of field, smooth engine-driven camera with zero handheld shake. Miami Vice City palette throughout: magenta, electric purple, cyan and orange neon, wet reflective asphalt, palm trees, art-deco facades. GAME HUD LOCK The GTA-style HUD is present, sharp and stable in every single frame of every segment, locked to screen space: a circular minimap in the bottom-left corner with dark base, glowing neon pink-purple roads and a white player arrow; a thin green health bar and blue armor bar above the minimap; five wanted stars in the top-right corner (lit or unlit as specified per segment); ammo counter in the bottom-right corner during weapon segments. Flat, clean, modern Rockstar UI language with a subtle neon glow. The HUD never disappears, never blurs, never moves with the camera. FORMAT MODE CONTROLLED MULTI-SHOT SEQUENCE — five segments joined by HARD CUTS only. No fades, no dissolves, no motion-blur transitions. Each segment opens with the character already fully in frame in the stated position — no empty frames, no delayed reveals. Real-time motion throughout, no slow motion. SEGMENT 1 — REAR CHASE CAM, CONVERTIBLE AT FULL SPEED Classic GTA third-person driving camera: positioned 4 meters behind the car, slightly above roof height, 47-degree field of view, smooth engine follow. The character from the reference image drives an open neon-purple convertible down a wide Vice City boulevard, neon hotel signs streaming past on both sides, palm silhouettes against the night sky. Wanted stars unlit. The car weaves once between two slower cars with real vehicle mass and suspension lean. Wet asphalt mirrors the neon strip. A brief held beat of pure speed. HARD CUT SEGMENT 2 — FRONT-FACING THIRD-PERSON CAM, FACE FULLY VISIBLE The in-game camera has been rotated a full 180 degrees to directly in front of the character, exactly like a player spinning the right stick: camera 2.5 meters ahead of him at chest height, facing him straight on, tracking backward as he walks toward the camera. The character from the reference image walks down the center of a neon-flooded street directly toward the lens, face fully lit and clearly visible, matching the reference image 100% — confident expression, eyes forward. Magenta and cyan signs blaze behind him, steam rising from a grate. His face is the focal point of the frame: centered, sharp, occupying the upper third. Natural walking physics — heel strike, weight shift, arms swinging. Wanted stars unlit. HARD CUT SEGMENT 3 — REAR SHOULDER CAM, SHOOTOUT BEHIND A CAR Third-person over-the-shoulder combat camera, 2 meters behind and slightly above, 47-degree field of view. The character from the reference image crouches behind the tail of a parked sports car, pops up over the trunk and fires a pistol three times at muzzle-flash-lit targets across the street. Real recoil: wrist kick, shoulder jolt, shell casings ejecting right and bouncing on the asphalt. Ammo counter reads 12/96 in the bottom-right. Two wanted stars lit in the top-right. Sparks burst off a wall across the street. Neon pink light rakes across the car body with every muzzle flash. HARD CUT SEGMENT 4 — FRONT-QUARTER ORBIT TO DEAD-ON FRONT VIEW, MOTORCYCLE The camera starts at the motorcycle’s front-right quarter and orbits smoothly to a dead-on front position, ending directly ahead of the bike at headlight height, tracking backward at matched speed — the character from the reference image rides a sport bike straight at the camera, no helmet, face clearly visible and 100% matching the reference image, hair whipped back by wind, jaw set. Four wanted stars lit and pulsing. Red-and-blue police strobes flash from behind him, staining the wet road. The bike’s headlight flares toward the lens; neon towers streak past on both edges of frame. Real bike physics: rider leaning low, suspension chatter over road seams. HARD CUT SEGMENT 5 — LOW REAR CAM, WALK-AWAY ON OCEAN DRIVE Final beat: third-person camera low and 3 meters behind. The character from the reference image walks calmly away from the camera along the Ocean Drive beach strip at dusk-into-night, art-deco neon hotels glowing pink and cyan on the left, dark ocean on the right, palms overhead. Wanted stars unlit again. He rolls his shoulders once and keeps walking into the neon glow as the segment holds for a long beat. Confident, iconic GTA end-frame energy. PHYSICS All motion obeys real mass and inertia: vehicle suspension compresses and rebounds, tires grip and slip at limits, recoil travels through the arm, casings bounce, clothing and hair react to wind and movement, footsteps carry weight transfer. Nothing floats, nothing teleports. LIGHTING Night-time Vice City throughout: neon signage is the primary light source — magenta, cyan, orange washes on skin, cars and wet asphalt. Police strobes act as dynamic red-blue accents in segments 3 and 4. In both front-facing segments the character’s face is cleanly lit by neon spill and readable at all times — never in silhouette, never in shadow. AUDIO Diegetic game audio only: engine roar, tire screech, gunshots with echo, police sirens, ocean ambience in the final segment, distant synthwave bass from a passing car. No narration, no dialogue, no captions. POSITIVE LOCKS Every segment contains only the one protagonist from the reference image. The HUD is present and identical in style across all five segments. The camera behaves like a player-controlled in-game camera at all times. Total duration 20 seconds, real-time pacing, five segments of roughly equal weight.

با افزودن عکس مرجع از یک سوژه زن و توصیف جزئیات ظاهری او در بخش Active Reference، می‌توانید ویدیو را با یک شخصیت زن بسازید.

پرامپت ادیت ویدیو با هوش مصنوعی

نمونه پرامپت ویدیو هوش مصنوعی برای تغییر سوژه از مرد به زن در پرامپت بالا:  

The protagonist is a young woman with long dark wavy hair, olive skin, brown eyes, and a confident expression. She wears a black leather jacket, white top, dark jeans and ankle boots. Keep her appearance consistent in every segment and camera angle.

۵۰ پرامپت آماده کپی و کاربردی برای ساخت عکس حرفه‌ای

۱۰. پرامپت تبدیل عکس به ویدیو هوش مصنوعی

• نام مدل: Seedance 2
• نسبت کادر خروجی: 16:9 
• مدت‌زمان: ۱۵ ثانیه

=== INPUT PARAMETERS === REF_WOMAN = @[Image 1](image_1)   → in this entire prompt set Image 1 is ALWAYS the woman REF_MAN   = @[Image 2](image_2)   → in this entire prompt set Image 2 is ALWAYS the man WOMAN_DESC = [fill in: age, build, hair length + texture + how it is worn, eye colour, one or two unique visible features] MAN_DESC   = [fill in: age, build, hair length + texture + how it is worn, facial hair or bare face, one or two unique visible features] ASPECT = 2.39:1 DURATION = 14 seconds === END INPUT PARAMETERS === SCENE CONTEXT A Persian chogan field in the late eleventh century, golden hour. Two mounted teams play. The woman is on the opposing team and she wins the point. The whole segment is fast, dusty and joyful, the only playful scene in the film. CHARACTER LOCK Scan every physical trait directly from the attached reference images and reproduce them exactly. Never hardcode, never idealise, never stylise, never age, never re-proportion. @[Image 1](image_1) is the woman, described by WOMAN_DESC. @[Image 2](image_2) is the man, described by MAN_DESC. Face geometry, bone structure, skin tone and finish, eye shape and colour, brow shape and density, nose shape, lip shape, hairline, hair colour and texture, and facial hair are taken from the references exactly as they appear. The two identities are never blended with each other and never swapped. Both faces stay bare with eyes fully visible in every shot, no eyewear, no mask, no face covering, no helmet. Pore-level skin realism, real catch-lights, natural blinking. RED THREAD LOCK Both characters wear a thin twisted red silk thread tied around the left wrist. Identical shade, identical thickness, identical knot in every shot. It is the only pure saturated red anywhere on either character. Whenever a left wrist or hand enters frame the thread is clearly visible. LOCATION MAP Foreground: dry sun-burnt grass and packed earth churned by hooves, heavy golden dust hanging at knee height. Midground: the open playing field with six to eight mounted riders on light fast horses, long-handled wooden mallets in hand, a small hard wooden ball in play. Background: low stone goal posts at the far end, a line of tall cypress trees, a mudbrick pavilion with a shaded arcade holding a small seated audience, copper sunset sky low behind them. Sun low and directly behind the far end of the field, backlighting all the dust. White balance 4800K, fixed for the whole scene. FIRST FRAME / BLOCKING First frame: camera planted at ground level, 20 cm above the earth, a wall of horse legs already thundering past frame left to right, backlit golden dust exploding across the lens. Riders are already at full speed in frame one. @[Image 1](image_1) and @[Image 2](image_2) are both in the pack, on opposing teams. FORMAT MODE Timed multishot, 14 seconds total, six segments. Cuts only at 2.5s, 5.0s, 7.0s, 9.0s and 11.5s, the camera does not cut on its own. OPTICS 0.0s to 2.5s: WS, 63°, extreme low angle. 2.5s to 5.0s: MS, 47°. 5.0s to 7.0s: WS, 63°. 7.0s to 9.0s: MS, 47°. 9.0s to 11.5s: MCU, 29°. 11.5s to 14.0s: WS, 63°. No drift mid-segment. Rectilinear, natural motion blur on hooves, mallets and dust. CAMERA 0.0s to 2.5s: locked off at ground level as the pack passes over and around it. 2.5s to 5.0s: stabilised lateral track running alongside @[Image 1](image_1) at 40 km/h, matched to her horse, camera at her chest height. 5.0s to 7.0s: fast upward tilt and pan following the struck ball against the sky, single continuous move. 7.0s to 9.0s: stabilised lateral track alongside @[Image 2](image_2) at 40 km/h at his chest height. 9.0s to 11.5s: handheld with 2 to 3 cm tremor, riding level with both horses as they converge. 11.5s to 14.0s: locked off at the field edge at chest height. Wide tonal latitude, soft highlight roll-off protecting the backlit dust. ACTION 0.0s to 2.5s — The pack gallops over and past the ground-level camera at 40 km/h. Clods of earth fly, backlit dust fills the frame, mallet shafts swing through. 2.5s HARD CUT 2.5s to 5.0s — Tracking with @[Image 1](image_1) at full gallop, mallet raised high, her scarf and hair streaming back. She rises in the stirrups, swings down hard and strikes the ball at 5.0s exactly. 5.0s HARD CUT 5.0s to 7.0s — The struck ball arcs high across the copper sky, the camera following it up and over in one continuous tilt and pan, cypress tops and low sun entering frame. 7.0s HARD CUT 7.0s to 9.0s — Tracking with @[Image 2](image_2) riding hard, head turned up following the ball, then turning down and across toward her, a competitive grin breaking across his face, teeth showing, dust on his cheek. 9.0s HARD CUT 9.0s to 11.5s — The two horses converge and cross at speed, knees almost touching, mallets held clear. Both riders turn their heads to each other for roughly half a second as they pass, then face forward again. Handheld tremor throughout. 11.5s HARD CUT 11.5s to 14.0s — Locked wide at the field edge. @[Image 1](image_1) reins in hard, the horse turning under her, and raises her mallet straight up in victory as the dust settles around her. Across the field @[Image 2](image_2) sits still on his horse, watching her, not moving. PERFORMANCE Both play to win and both are enjoying it. Her effort reads through the whole body on the swing, jaw set, then breaks into open delight at the end. His grin is real and unguarded, the only time he smiles this widely. Sweat sheen, dust on skin, wind-flushed cheeks, visible hard breathing, clear catch-lights. PHYSICS Horses carry full mass and inertia, hooves throw earth in clumps, tack and stirrups bounce, manes and tails move with the gallop. The mallet swing carries full body torque and follow-through. The ball leaves the mallet with real impulse and arcs on a true ballistic path. Dust rises, drifts in the low wind and settles. Contact shadows anchor every horse to the ground. LIGHTING Single low golden hour sun directly behind the far end of the field, one direction in every cut. Every rider is rim-lit and every dust cloud is backlit and glowing. Faces are filled by warm bounce off the pale dry earth. Long hard shadows stretch toward camera. No artificial fill. COLOR GRADE Golden dust and copper sky dominant, dry olive and burnt green in the grass, deep cypress green holding the darkest midtones, warm saturated skin. High warmth and high atmosphere with the strongest saturation in the film. The red wrist threads sit inside the warm palette and stay readable. Rich blacks in the pavilion arcade. WARDROBE @[Image 1](image_1): fitted knee-length riding coat in undyed linen over narrow trousers, a long light scarf wrapped once at the neck and streaming behind her with the face fully uncovered, soft riding boots, sleeves ending above the wrist so the red thread stays visible. @[Image 2](image_2): fitted riding tunic cinched with a leather belt, narrow trousers, soft riding boots, a light cloth band at the forehead leaving the face and eyes fully uncovered, sleeves ending above the wrist so the red thread stays visible. Other riders: eleventh century Persian riding dress in undyed linen and leather, two team colours in muted ochre and muted indigo only, never red. RANDOMISATION Every other rider’s face, every horse’s colour and markings, the audience in the pavilion and every incidental prop are invented by the model itself and come out different on every run. Do not repeat a fixed set. AUDIO Thundering hooves dominant throughout, sharp wooden crack of mallet on ball at 5.0s, riders calling to each other wordlessly, tack and leather creaking, wind, a distant drum from the pavilion, a rising cheer at 12.0s. No dialogue. STYLE Photoreal live-action historical cinema, 8K detail, fine natural film grain, anamorphic widescreen character, real-time speed in every segment. OUTPUT SETTINGS 14 seconds, 2.39:1, 8K photoreal, real-time speed in every segment. POSITIVE LOCKS @[Image 1](image_1) is the woman and @[Image 2](image_2) is the man, never reversed. Both faces bare, eyes uncovered, no helmet and no face covering at any point, identical to the references in every shot. Red thread on the left wrist of both in every shot where a wrist is in frame. Play direction holds: the ball and the pack always move left to right. The sun stays low and behind the far end of the field in every cut, 4800K held across all cuts. Nothing in this scene is red except the two wrist threads. The woman wins the point, not the man. Cuts only at 2.5s, 5.0s, 7.0s, 9.0s and 11.5s. HANDOFF FRAME Last frame: wide, dust settling, her mallet raised, him still across the field. Editorially, the ball arc in segment three is the intended match-cut partner for the dive arc that opens S04.

هنگام تولید کلیپ با پرامپت ساخت ویدیو از عکس، از تصاویر مرجع باکیفیت و ابعاد یکسان استفاده کنید. پیش از آپلود تصاویر، با مراجعه به بخش «استایل‌های آماده» در هوشا، مدل تصاویر مرجع در استایل موردنظر را بررسی کنید و از سوژه خود با زاویه‌هایی مشابه تصاویر نمونه عکاسی کنید. برای ساخت پرتره از زوایای مختلف نیز می‌توانید تنها یک عکس واضح و تمام‌رخ از سوژه را در سرویس ساخت عکس هوشا آپلود کنید.

کدام هوش مصنوعی برای ساخت ویدیو با پرامپت مناسب‌تر است؟

پرامپت ویدیو چت جی پی تی، پرامپت ویدیو گروک و پرامپت ویدیو جمنای شباهت زیادی به یکدیگر دارند و با تغییرات جزئی، مانند کاهش تعداد کاراکترهای متن یا افزودن و حذف بخش‌های مربوط به ساخت آهنگ پس‌زمینه، می‌توان از آن‌ها در ابزارهای مختلف استفاده کرد. هنگام انتخاب ابزار نیز، علاوه بر امکانات ساخت ویدیو، به قابلیت‌های آن در نگارش و اجرای پرامپت توجه کنید:

• محدودیت تعداد کاراکتر متن 
• گزینه بهبود پرامپت
• امکان افزودن ویدیوی رفرنس علاوه بر پرامپت متنی

برای ساخت پرامپت ویدیو از صفر نیز، ابتدا متن پرامپت را با چت جی بی تی فارسی بنویسید و اصلاح کنید. سپس آن را در چند ابزار مختلف امتحان کنید و با مقایسه خروجی‌ها، مدلی را انتخاب کنید که نتیجه مورد نظر شما را ارائه می‌دهد.

چطور پرامپت ویدیو را برای نتیجه بهتر شخصی‌سازی کنیم؟

برای پرامپت نویسی ویدیو می‌توانید از هوش مصنوعی آنلاین فارسی کمک بگیرید و سپس آن را ترجمه کنید. رعایت ۴ نکته زیر هنگام نوشتن پرامپت، به دریافت خروجی دقیق‌تر کمک می‌کند:

1. موضوع و اکشن (کاری که سوژه انجام می‌دهد) را دقیق توضیح دهید. در پرامپت تبدیل عکس به ویدیو، اکشن‌ها به‌صورت زمان‌بندی‌شده و با جزئیات توصیف شده‌اند.
2. حرکت دوربین را مشخص کنید؛ مثل حرکت به چپ و راست دوربین (Pan Left / Right)، زوم یا دنبال کردن سوژه (Tracking). برای مثال، در پرامپت ویدیو استادیوم، حرکت دوربین و زاویه‌های مختلف مشخص شده است.
3. نورپردازی و سبک تصویری را مشخص کنید. عباراتی مثل «Bright daylight» در پرامپت ویدیو طبیعت یا «Cinematic» در پرامپت ویدیو ترند، به تعیین نور و حال‌وهوای صحنه کمک می‌کنند. تکرار این جزئیات در بخش‌های مرتبط متن نیز باعث می‌شود سبک تصویر هماهنگ بماند.
4. مدت‌زمان و نسبت تصویر را متناسب با پلتفرم مقصد انتخاب کنید. برای ریلز و استوری اینستاگرام نسبت عمودی و برای یوتیوب نسبت افقی مناسب است.

ساخت ویدیو و دسترسی به پرامپت آماده در هوشا

برای پیدا کردن پرامپت آماده برای ساخت فیلم با هوش مصنوعی در هوشا، وارد بخش «ساخت ویدیو» شوید و روی تب «استایل‌های آماده» بزنید. در این بخش، فهرستی از پرامپت‌های آماده و دسته‌بندی‌شده را می‌بینید. پرامپت‌های این مقاله نیز از بخش استایل آماده در هوشا در سبک‌های مختلف واقع‌گرا، انیمیشن، طنز و … ساخته شده‌اند. 

البته در هوشا این امکان را دارید که پرامپت خود را کاملاً از صفر بنویسید و تمام جزئیات فنی از انتخاب مدل هوش مصنوعی گرفته تا تعیین نسبت تصویر، مدت‌زمان و کیفیت خروجی را طبق نیاز خود تنظیم کنید.

مهم‌ترین مزایای ساخت ویدیو در پلتفرم هوشا (Hoosha)

هوشا یک هوش مصنوعی فارسی همه کاره با بیش از ۲ میلیون کاربر است که امکان استفاده از پرامپت ساخت ویدیو را در کنار مدل‌ها و ابزارهای مختلف تولید محتوا، از جمله موزیک، متن و تصویر، فراهم می‌کند.

در هوشا، علاوه بر ۱۰ مدل‌ مختلف تولید ویدیو از جمله GrokImagine 1.5، Seedance، Veo، Kling، Runway و Google Omni مجموعه‌ای از استایل‌های آماده نیز برای ایده گرفتن یا ساخت ویدیو وجود دارد. مهم‌ترین مزایای هوشا برای ساخت ویدیو عبارت‌اند از:

• محیط کاربری فارسی و ساده با دسترسی از طریق وب‌سایت، وب‌اپلیکیشن، ربات تلگرام و اپلیکیشن رسمی اندروید
• پرامپت‌ها و قالب‌های آماده برای ساخت سریع‌تر محتوا 
• دسترسی بدون نیاز به VPN، با درگاه پرداخت ریالی و امکان پرداخت از طریق اسنپ‌پی

برای امتحان استایل‌های آماده ساخت ویدیو در ابزارهای پیشرفته هوش مصنوعی، از طریق لینک زیر وارد صفحه هوش مصنوعی ساخت ویدیو در هوشا شوید.

مقایسه قابلیت‌های مدل ساخت ویدیو در هوشا

در جدول زیر می‌توانید ویژگی‌های اصلی مدل‌های استفاده‌شده در ۱۰ پرامپت این مقاله و قابلیت‌های در دسترس آن‌ها در هوشا را مشاهده کنید.

مدل ویدیو	ویژگی شاخص	درک پرامپت فارسی	مدت‌زمان / ورودی	قابلیت ساخت موزیک پس‌زمینه
Grok Imagine 1.5	ویدیوهای واقع‌گرایانه و درک بهتر پرامپت	✓	۳ تا ۱۵ ثانیه	✓
Kling 3.0 Turbo	تصویرسازی واقع‌گرایانه و انیمیشن‌سازی	✕	۳ تا ۱۵ ثانیه	✓
Seedance 2.5	پشتیبانی از ویدیوی رفرنس؛ قابلیت آپلود چند رفرنس همزمان	✕	۴ تا ۳۰ ثانیه	✓
Seedance 2	تیزرهای سینمایی و شات‌های متنوع دوربین	✕	۴ تا ۱۵ ثانیه	✓
Kling Motion	انتقال حرکت، ژست و حالت چهره از ویدیوی مرجع	✕	۳ تا ۳۰ ثانیه؛ بر اساس مدت ویدیوی مرجع	✕
Veo 3.1	ویدیوهای سینمایی با صدای طبیعی	✓	۸ ثانیه	✓

جمع‌بندی

در این مقاله، ۱۰ پرامپت ویدیو با هوش مصنوعی را برای موضوعات متنوع، از طبیعت و ورزش تا محتوای ترند و سرگرمی، آماده کرده‌ایم. هر پرامپت همراه با نمونه خروجی و تنظیمات پیشنهادی ارائه شده تا بتوانید ساختار آن را بهتر بررسی و متناسب با ایده خود شخصی‌سازی کنید.

برای ساخت ویدیو نیازی به نوشتن متن پرامپت از صفر ندارید. می‌توانید نمونه‌های آماده را در هوشا (Hoosha) اجرا کنید، نتیجه را ببینید و با تغییر بخش‌های مختلف، پرامپت ویدیو با هوش مصنوعی را متناسب با نیاز خود تنظیم کنید.

نویسنده: سپیده بزازی

• منبع ۱
• منبع ۲
• منبع ۳