Guide
Replicate the Riverflow VELLA Carousel
A copy-ready prompt guide for recreating Riverflow's VELLA carousel, including six still-image concepts, one Nano Banana Pro retouch, and three approved animations.
- Guides
- Brand Creative Workflows

Examples
Scenes from the Riverflow library

Frame 1: Rome Installation

Frame 2: FRESA Slide

Frame 3: Hooded Beauty

Frame 4: LIMA Bridge

Frame 5: MORA Swing

Frame 6: Strawberry Cover
The process
Follow the source workflow in order: product first, then image, then video.
Riverflow workflow
Product first, then image, then video
Bring the finished garment, label design, check pattern, fringe detail, and colourways. Create one product sheet showing the garment open, folded, the woven-label macro, and every colourway, then lock it as the shared reference. The source does not provide a product-sheet generation prompt.
Product
Create and lock the product sheet
Use finished product decisions rather than placeholders. The sheet keeps the check scale, woven label, fringe, and colourways consistent across the carousel.
Image
Generate the still
Open Riverflow's image generator, choose Riverflow 2.5, attach the product sheet and the other numbered references required by the full prompt, paste the prompt, and generate.
Video
Animate the approved still
Open Riverflow's video generator, choose Seedance 2.5, use the approved still as the start frame, add the product sheet as the second reference, paste the matching motion prompt, and generate for six or seven seconds in vertical 3:4.
Six Riverflow 2.5 prompts
Visual playbook
The six-frame VELLA carousel
Use the complete prompt for each frame, keep its numbered reference roles intact, and output every still at exactly 3456 × 4320 pixels in vertical 4:5. Frame 2 adds a Nano Banana Pro retouch after the Riverflow 2.5 generation.

1. Rome Installation
A real out-of-home installation in historic Rome, with an enormous FRESA cape spilling from the billboard into the street.
Use when: Use Riverflow 2.5 and output 3456 × 4320 pixels in vertical 4:5. Follow the first, second, and third reference roles exactly as written.
Original prompt
USE CASE: photograph of a real out-of-home advertising installation on a city street — a documentary-style street photo of a giant 3D billboard, the kind a brand would publish as a campaign case study.
STYLE: recreate the FIRST attached reference photograph exactly — the same street-photography framing, the same enormous framed billboard mounted high on the corner of a stone building with the advertised textile spilling out of the poster as a real three-dimensional object and cascading down the facade over the street, the same natural late-afternoon daylight, the same crisp realistic photographic rendering. Not a rendering, not a mockup overlay — a real photograph of a real installation.
PRESERVE FROM THE FIRST REFERENCE (do not alter): the camera position standing on the pavement looking slightly upward along a narrow street; the ornate stone building occupying the left third with the huge framed portrait billboard mounted high on its corner; the enormous piece of check fabric emerging from the bottom of the billboard as a real physical object, swelling into full three-dimensional volume, sweeping down and out across the facade and billowing over the street with visible weight and catenary curve; the pedestrian seen from behind walking away down the pavement below, carrying an open umbrella made of the same check; the café terrace on the left with bistro chairs, small tables and square planters with clipped topiary; the tall iron street lamp; the parked bicycles at the right kerb; scattered pedestrians in the mid-distance; the street receding in perspective with buildings on both sides; the clear blue sky in the upper right; the warm raking late-afternoon sun grazing the stonework and the long soft shadows it casts.
CHANGE: the city, the brand and the fabric, as described below.
SCENE: a narrow street in the historic centre of ROME. Travertine and burnt-ochre rendered facades, tall shuttered wooden windows with peeling paint, a wrought-iron balcony with trailing ivy, a small marble cornice, terracotta roofline in the distance. A vintage Vespa parked at the kerb. The café terrace is a trattoria terrace with cane bistro chairs and small marble-topped tables. Warm golden late-afternoon Roman light, deep clean shadows, clear blue sky.
THE BILLBOARD: a huge framed poster mounted high on the building corner, with a wide pale border. At the top of the poster, the wordmark “VELLA” set large, letterspaced, in near-black geometric grotesque capitals. Below it, a waist-up studio portrait of the SAME young woman from the third attached reference image — identical person, same face, same features: early twenties, long straw-blonde hair, fair skin with faint freckles, pale eyes, calm direct gaze into camera — wearing the VELLA fringed brushed-mohair cape draped around her shoulders and neck, photographed against a plain pale studio ground.
THE GIANT FABRIC: the very same cape continues out of the poster into real three-dimensional space — an enormous fringed brushed-mohair CAPE in the “FRESA” colourway from the second attached reference sheet: fuchsia, coral, dusty rose and cream oversized block check, long-pile fluffy mohair, with long hand-twisted fringe. It is vividly, intensely saturated, blazing against the ochre Roman stone — this colour contrast is the whole point of the image. It falls from the poster with real fabric weight, folds and volume, wraps around the building corner and billows out over the street, the low sun catching its pile and making the fibres glow at the edges. Fringe hangs from its short end.
THE UMBRELLA: the pedestrian’s open umbrella is made from the same FRESA check, a small echo of the giant fabric above.
CAMERA: full-frame digital body, 28mm lens at f/5.6, handheld at eye level, tilted slightly up. Deep focus across the street, with only the very nearest café chairs at the left edge slightly soft. Natural perspective convergence in the architecture.
TEXT (EXACT): “VELLA” — the only legible text in the image is the wordmark at the top of the billboard, sharp, correctly spelled, correctly letterspaced.
PALETTE: #C9A227 / #D9C29A / #E8336E / #F4633C / #EFB7BC / #F6EDE4 / #6FA8DC / #3A3226
CONSTRAINTS: no graphic overlays of any kind, no logos beyond the billboard wordmark, no watermarks, no hashtags. Do not depict any real, famous or recognisable person, and do not reproduce any existing real fashion brand’s name, monogram or signature check pattern anywhere. The fabric must read as a wide fringed cape panel, never a long narrow scarf.
OUTPUT: Output in exactly 3456px x 4320px (4:5 portrait vertical format).
Negative prompt: existing luxury brand logo, real brand name, trademarked check pattern, long narrow scarf, flat pasted-on fabric, CGI render look, text overlay, watermark, misspelled wordmark, garbled lettering, different model, deformed hands, extra fingers, duplicate subject, plastic skin, AI-smoothed skin, oversaturated sky, HDR, lens flare, celebrity likeness, recognisable real person.

2. FRESA Slide
A giant FRESA cape becomes a broad playground slide in a cream-to-strawberry studio gradient.
Use when: Generate the still with Riverflow 2.5 at 3456 × 4320 pixels in vertical 4:5, then apply the exact Nano Banana Pro motion retouch below.
Original prompt
USE CASE: high-end fashion campaign photograph for a knitwear label — surreal-scale key visual, clean and publishable.
STYLE: recreate the FIRST attached reference photograph exactly — the same playful giant-textile-as-landscape idea, the same crisp cut-out subject and fabric floating in a seamless studio gradient, the same joyful energy, the same saturated but controlled editorial colour.
PRESERVE FROM THE FIRST REFERENCE (do not alter): the overall composition and crop, the enormous ramp of woven fabric sweeping diagonally from the upper right of the frame down toward the lower left, the long twisted fringe hanging off the ramp’s lower end, the single figure seated on the ramp mid-slide with knees bent and legs forward, arms flung out for balance, hair streaming with the movement, the open laughing expression, the large woven brand label lying flat on the fabric surface below the figure, the floating cut-out treatment with no floor and no horizon, and the soft even large-source key light.
SCENE: seamless studio sweep with a smooth gradient from warm cream at the top to soft strawberry pink at the bottom. No floor line, no horizon, no props.
SUBJECT: the SAME young woman from the third attached reference image — early twenties, long straw-blonde hair, fair skin with visible natural texture and faint freckles, pale eyes, soft neutral makeup. Caught mid-slide, seated on the giant fabric ramp, sliding down toward the lower left, both arms flung wide for balance, head tilted back, laughing openly, hair streaming up and back with the speed.
THE GIANT GARMENT — HERO: an enormous oversized fringed brushed-mohair CAPE in the “FRESA” colourway — fuchsia, coral, dusty rose and cream oversized block check, long-pile fluffy mohair, scaled up to the size of a playground slide.
CRITICAL GEOMETRY: it must read unmistakably as a WIDE RECTANGULAR CAPE PANEL, not a long narrow scarf. The ramp is BROAD, with its full oversized check grid reading many squares across the width. The long hand-twisted FRINGE appears only along the SHORT END of the panel; the two long sides are clean selvedge edges with NO fringe. The white woven rectangular label reading “VELLA” lies flat on the fabric surface below the figure, scaled up huge, fully legible.
WARDROBE: a fitted long-sleeve cream ribbed cotton top with the wordmark “VELLA” printed small and letterspaced across the chest; wide-leg sage-green trousers in a fine lurex-flecked weave; cognac-brown shearling-lined clogs.
CAMERA: medium-format digital body, 50mm equivalent lens at f/2.8. Focus locked on the woman; the far upper end of the ramp falls clearly out of focus.
TEXT (EXACT): “VELLA” — appears exactly twice: large on the giant woven fabric label, and small on the chest of the ribbed top.
PALETTE: #F6EDE4 / #E8336E / #F4633C / #EFB7BC / #F2EFEA / #8CA07A / #7A4B2A / #161616
CONSTRAINTS: no graphic overlays, no logos, no watermarks. The giant fabric must be a wide cape panel — do not render it as a long narrow scarf and do not put fringe along its long sides.
OUTPUT: Output in exactly 3456px x 4320px (4:5 portrait vertical format).
Negative prompt: long narrow scarf, fringe along the long edges, wrong check scale, missing woven label, garbled lettering, deformed hands, plastic skin, oversaturated, HDR, recognisable real person.

3. Hooded Beauty
An extreme overhead fisheye portrait with the FRESA cape cowled around the subject's face.
Use when: Use Riverflow 2.5 and output 3456 × 4320 pixels in vertical 4:5.
Original prompt
USE CASE: high-end fashion campaign photograph for a knitwear label — beauty-led key visual, clean and publishable.
STYLE: recreate the FIRST attached reference photograph exactly — the same extreme wide-angle look-down beauty portrait, the same off-white studio void, the same crisp saturated editorial colour, the same intimate upward gaze.
PRESERVE FROM THE FIRST REFERENCE (do not alter): the extreme high-angle camera position looking straight down at the subject, the subject’s face tilted up toward the lens and dominating the upper half of the frame, the dramatic foreshortening of the body receding away from camera toward the bottom of the frame, the tiny distant shoes at the very bottom edge, one arm reaching diagonally up and out toward the upper-left corner, the crisp cut-out subject floating in an empty pale studio void, and the soft even wraparound key light.
SCENE: empty seamless off-white studio void (#F2EFEA), very faintly warming toward pale strawberry pink in the lower corners.
SUBJECT: the SAME young woman from the third attached reference image — early twenties, long straw-blonde hair, fair skin with faint freckles across the nose, soft neutral makeup, pale eyes. She tilts her chin and eyes UP toward the lens directly above her, lips slightly parted, a calm magnetic expression. Seen from directly above: her head and face fill the upper portion of the frame close to the lens and are TACK-SHARP; her shoulders and torso recede steeply below her; her legs taper away in strong perspective; her small cognac-brown leather boots sit tiny at the very bottom edge of the frame.
WARDROBE: the oversized fringed brushed-mohair CAPE in the “FRESA” colourway — fuchsia, coral, dusty rose and cream oversized block check, long-pile fluffy mohair — wrapped as a HOOD, cowled softly around her head and framing her face, then coiled around her neck and shoulders. One long fringed end is caught mid-air, sweeping diagonally out toward the upper-left corner, its hand-twisted fringe splayed and backlit. The small white woven rectangular label reading “VELLA” is clearly visible on that flying end, close to her face and fully legible. Over the cape’s shoulders, a glossy translucent coral raincoat catching wet-look specular highlights.
CAMERA: full-frame body, 15mm fisheye lens at f/2.8, held directly overhead. Pronounced barrel distortion. Focus plane locked on the eyes: the face, the hood of mohair and the woven label are tack-sharp with visible fibre detail, while the legs blur and the boots render as soft creamy bokeh.
TEXT (EXACT): “VELLA” — the only text in the image, on the small woven cape label.
PALETTE: #F2EFEA / #E8336E / #F4633C / #EFB7BC / #F6EDE4 / #7A4B2A / #161616
CONSTRAINTS: no graphic overlays, no watermarks. Do not turn the cape into a narrow scarf — it must read as a wide fringed mohair cape even when wrapped as a hood. Do not render the whole frame equally sharp.
OUTPUT: Output in exactly 3456px x 4320px (4:5 portrait vertical format).
Negative prompt: flat all-in-focus rendering, sharp shoes at the bottom, rectilinear undistorted lens, narrow scarf instead of wide fringed cape, missing woven label, deformed hands, plastic skin, oversaturated, recognisable real person.

4. LIMA Bridge
A broad LIMA cape panel forms a suspended bridge through a cream-to-lime studio void.
Use when: Use Riverflow 2.5 and output 3456 × 4320 pixels in vertical 4:5.
Original prompt
USE CASE: high-end fashion campaign photograph for a knitwear label — surreal-scale brand hero visual, clean and publishable.
STYLE: contemporary studio fashion campaign with playful scale distortion — a single figure inside a world made of one enormous piece of knitwear, crisply cut out and floating in a seamless colour-gradient void.
SCENE: seamless studio void with a smooth vertical gradient from warm cream at the top to soft lime green at the bottom.
THE GIANT GARMENT — THE HERO, AND ITS SHAPE MUST BE UNMISTAKABLE: an enormous oversized fringed brushed-mohair CAPE in the “LIMA” colourway — lime green, sage, butter yellow and cream oversized block check. It is scaled up enormously and suspended taut across the scene like a BRIDGE or a flat hammock, running from the lower left of the frame back and away toward the right.
READABILITY OF THE PANEL IS THE POINT: the camera is slightly ABOVE the panel, at a three-quarter angle, looking down along its length, so that the NEAR long edge, the FAR long edge and the whole flat top surface between them are all clearly visible in perspective. The check grid reads as a proper woven grid: many oversized squares across the width AND many squares running away along the length. The panel is BROAD: its width is greater than the standing woman’s height. Long hand-twisted FRINGE hangs only from the SHORT END, at the right of the frame; the two LONG edges are clean straight selvedge with absolutely NO fringe.
THE LABEL: the white woven rectangular label reading “VELLA” lies flat on the fabric surface in the near foreground, scaled up huge, tack-sharp and the single most legible element in the image.
SUBJECT: the SAME young woman from the reference image — early twenties, long straw-blonde hair, fair skin with faint freckles, pale eyes. She walks along the top of the giant panel toward the camera, both arms opened wide and low for balance, weight sinking slightly into the soft cloth, head tilted down watching her step, a private delighted half-smile.
WARDROBE: a sheer terracotta-rust organza midi dress with soft puffed short sleeves, worn over a fitted cream ribbed top; cognac-brown leather knee boots.
CAMERA: medium-format digital body, 85mm equivalent lens at f/2.8, positioned slightly above the plane of the fabric. Focus locked on the giant woven label and the walking figure; the far end of the panel falls clearly out of focus, dissolving into creamy bokeh.
TEXT (EXACT): “VELLA” — the only text in the image, on the giant woven label.
PALETTE: #F6EDE4 / #B8D24A / #8CA07A / #F0D278 / #E6F0C8 / #B4552A / #7A4B2A / #161616
CONSTRAINTS: the giant fabric must read as one flat wide rectangular cape panel seen in perspective — it must NOT curve into an ambiguous U-shaped sling, must NOT be a long narrow scarf, and must NOT have fringe along its long edges.
OUTPUT: Output in exactly 3456px x 4320px (4:5 portrait vertical format).
Negative prompt: U-shaped sling, ambiguous fabric shape, unreadable fabric geometry, long narrow scarf, fringe along the long edges, missing woven label, deformed hands, feet floating above the fabric, oversaturated, recognisable real person.

5. MORA Swing
A human-scale FRESA cape and giant MORA sweeps frame a rope swing suspended in a cream-to-lavender void.
Use when: Use Riverflow 2.5 and output 3456 × 4320 pixels in vertical 4:5.
Original prompt
USE CASE: high-end fashion campaign photograph for a knitwear label — surreal-scale key visual, clean and publishable.
STYLE: dreamlike giant-textile world, seamless cream-to-lavender gradient void, soft even large-source key light, saturated but controlled editorial colour.
COMPOSITION: a full-body figure seated on a simple rope swing suspended in mid-air near the centre of the frame, one hand gripping each rope, arms raised, legs extended forward, weightless, soft delighted smile, hair drifting. One enormous sweep of check fabric arcs across the TOP of the frame and a second sweeps across the BOTTOM, both cut off by the frame edges with long twisted fringe at their short ends. A huge woven brand label lies on the lower sweep.
THE SWING ROPES — CRITICAL: two plain natural-fibre ropes rise from the simple wooden seat and travel STRAIGHT UP, cleanly and separately, exiting the image at the TOP EDGE of the frame. They must NEVER pass behind, cross, overlap, touch, disappear into or visually merge with the giant mohair sweeps.
TWO SCALES OF THE SAME PRODUCT — the core idea of the image: 1. GIANT — two enormous sweeps of the cape in the “MORA” colourway: deep blue, lavender, mauve and cream oversized block check. The woven “VELLA” label lies on the lower sweep, scaled up huge, fully legible. 2. HUMAN SCALE — the very same cape design WORN by the woman at normal human size, in the “FRESA” colourway: fuchsia, coral, dusty rose and cream, thrown over both shoulders with its own small woven label near the front hem.
CRITICAL GEOMETRY (both scales): the fabric reads as a WIDE RECTANGULAR CAPE PANEL, never a long narrow scarf. Hand-twisted FRINGE only along the SHORT ENDS; the two long sides are clean selvedge with NO fringe.
WARDROBE: a fitted long-sleeve cream ribbed cotton top with the wordmark “VELLA” small at the chest; wide-leg BUTTER-YELLOW trousers in a fine lurex-flecked weave; cognac-brown shearling-lined clogs.
CAMERA: medium-format digital body, 80mm equivalent lens at f/2. Focus locked precisely on the woman and the FRESA cape she wears. VERY STRONG optical depth-of-field separation — both giant sweeps heavily out of focus, their check squares dissolved into broad soft blocks of colour, though the huge woven label stays readable through the softness.
TEXT (EXACT): “VELLA” — appears exactly three times: large on the giant woven fabric label, small on the worn cape’s label, and small on the chest of the ribbed top.
PALETTE: #F6EDE4 / #3B5BA9 / #8E86C8 / #A8629B / #C9CFE8 / #E8336E / #F4633C / #EFB7BC / #F0D278 / #7A4B2A / #161616
CONSTRAINTS: neither cape may be rendered as a long narrow scarf or have fringe along its long sides. The worn FRESA cape must stay clearly distinct in colour from the giant MORA sweeps. The ropes must exit the top edge of the frame in clear open space and must not intersect the fabric.
OUTPUT: Output in exactly 3456px x 4320px (4:5 portrait vertical format).
Negative prompt: ropes merging with the mohair, ropes hidden behind the giant cape, sharp upper fabric sweep, both capes identical colour, missing woven label, deformed hands, oversaturated, recognisable real person.

6. Strawberry Cover
The FRESA cape is worn at human scale atop a photorealistic giant strawberry in a warm-cream-to-red sweep.
Use when: Use Riverflow 2.5 and output 3456 × 4320 pixels in vertical 4:5.
Original prompt
USE CASE: high-end fashion campaign photograph for a knitwear label — season key visual, clean and publishable.
STYLE: recreate the FIRST attached reference photograph exactly — the same surreal-scale studio fashion look, the same crisp cut-out subject floating against a seamless colour-gradient sweep, the same saturated but controlled colour, the same soft beauty lighting.
PRESERVE FROM THE REFERENCE (do not alter): the overall composition and crop, the subject’s scale and placement in frame, the seated pose on top of a giant piece of fruit, the raised arm gripping the fruit’s stem which rises vertically out of the top of the frame, the legs hanging with one knee bent, the floating cut-out treatment with no floor and no horizon, the seamless vertical colour gradient background, and the soft even large-source key light.
SCENE: seamless studio sweep with a smooth vertical gradient running from warm cream at the top to deep strawberry red at the bottom.
SUBJECT: a young woman in her twenties with long straw-blonde hair falling loose past her shoulders, fair skin with visible natural texture and faint freckles, soft neutral makeup, a calm confident half-smile. She sits balanced on top of an enormous, photorealistic GIANT STRAWBERRY — deep glossy red, dimpled with tiny golden seeds, its green calyx spread beneath her, and a thick green stem rising vertically out of the top of the frame, which she grips with one raised hand.
WARDROBE (render with great care, this is the product): the oversized fringed brushed-mohair CAPE in the “FRESA” colourway — fuchsia, coral, dusty rose and cream oversized block check, long-pile fluffy mohair, long hand-twisted fringe — worn at normal human scale, draped over both shoulders and falling open down her back and one arm. The small white woven rectangular label reading “VELLA” is visible stitched near the front hem. A fitted long-sleeve cream ribbed cotton top underneath, with the wordmark “VELLA” printed across the chest in small letterspaced near-black geometric grotesque capitals. Wide-leg sage-green trousers in a fine lurex-flecked weave. Cognac-brown shearling-lined clogs.
CAMERA: medium-format digital body, 85mm equivalent lens at f/4, full-length framing, subject tack-sharp edge to edge.
PALETTE: #F6EDE4 / #E8336E / #F4633C / #EFB7BC / #C2183C / #8CA07A / #7A4B2A / #161616
CONSTRAINTS: the ONLY text anywhere in the image is “VELLA” on the chest of the top and on the small woven cape label. Do not turn the cape into a narrow scarf — it must read as a wide fringed cape.
OUTPUT: Output in exactly 3456px x 4320px (4:5 portrait vertical format).
Negative prompt: narrow scarf instead of wide fringed cape, wrong colourway, missing woven label, orange instead of strawberry, deformed hands, distorted typography on the t-shirt, oversaturated, recognisable real person.
Create it in Riverflow
Frame 2: Nano Banana Pro motion retouch
Use this structure to turn the strategy into a specific creative brief that keeps the product accurate and the scene useful.
- 1
Step 1 — Riverflow 2.5
Generate the 4:5 FRESA Slide still with the exact Frame 2 prompt above.
- 2
Step 2 — Nano Banana Pro
Apply the exact retouch below, then confirm the composition and woven label remain unchanged.
Example prompt
Motion retouch (Nano Banana Pro, applied after generation): add a subtle directional motion blur to the woman so she reads as sliding at speed toward the lower left — body, arms and hair softly streaked along that diagonal — while keeping her face, the “VELLA” chest print, the fabric, the woven label and the background pixel-identical and sharp. Note: this model tends to re-render the whole frame rather than edit it — check the composition and label came back unchanged before accepting the result.
Three Seedance 2.5 animations
Use the matching approved still as the first frame and the product sheet as the second reference. The source specifies vertical 3:4 for these animations, while the image prompts specify vertical 4:5.
Model examples
The approved VELLA animations
The clips use Seedance 2.5 with the recorded six- or seven-second duration in vertical 3:4.
Video
Rome Installation
The FRESA installation rolls through a slow, weighty wave as pedestrians move beneath it.
Original prompt
One continuous handheld shot at standing eye level, like an amateur phone recording — small natural reframing corrections and mild exposure breathing, no visible phone or camera. The oversized VELLA cape moves with a slow, heavy wind, as if it has real textile mass; the fringe lags behind the main fabric and settles a fraction later. Two or three pedestrians walk underneath the draped fabric at normal pace, giving scale. The camera makes a tiny natural drift while the cape rolls through one more soft wave, with believable inertia and weight — never floating or moving like paper. The woman stays calm and still, holding the fashion-poster elegance of the original image. The motion settles into a dreamy final hero moment: the cape lowers back into place with a last gentle flutter in the fringe, and one passerby crosses below the fabric to emphasise scale. Audio: soft mohair rustle, cloth breathing in the wind, stone sidewalk footsteps, distant street chatter, one light umbrella rustle, quiet city ambience — no music, no dialogue, no subtitles.
Video
Hooded Beauty
The subject turns the woven VELLA label toward the lens, then hides behind the cape with a shy laugh.
Original prompt
One continuous handheld shot, casual and intimate, eye level, small natural drift, no visible device. She is already facing camera with the cape draped over her shoulders. She reaches with both hands, gathers the cape, and turns the woven VELLA label toward the lens so it is clearly readable, the fringe lagging slightly behind the main fabric with believable weight. She then gives a timid, innocent laugh, lowers her chin, and hides part of her mouth and cheeks behind the cape in a playful self-conscious gesture, her eyes peeking over the fabric while the cape settles with a final soft flutter. One continuous woman, one cape, one label, one camera angle throughout — no wardrobe changes, no face warping. Audio: soft mohair rustle, a small shy breath, gentle fabric movement, light cloth flutter, quiet room tone — no music, no dialogue, no subtitles.
Video
MORA Swing
The swing and FRESA cape settle naturally while the giant MORA sweep stays readable below.
Original prompt
Ultra-realistic single continuous shot, camera mostly locked with a very gentle slow push-in and tiny natural handheld breathing — no whip pans, no orbit, no crash zoom. The swing is already in motion with a very small gentle sway; the worn FRESA cape moves softly with the wind, the fringe lifting and falling naturally, while the woman keeps both hands on the ropes and gives one quiet blink and a relaxed micro-smile. The camera makes a barely perceptible push-in while the swing eases forward and back once more; her hair shifts slightly across her cheek, and the cape drapes and ripples in a physically believable way, the lower giant MORA sweep staying readable through its defocus. The motion settles toward centre: the fringe comes to rest with a final soft flutter, her grip relaxes a little, and she holds a calm, effortless pose that feels like a finished hero frame. Audio: soft rope creak, light swing sway, airy mohair rustle, gentle fabric flutter, a small playful whoosh as the swing settles — no music, no dialogue, no subtitles.
Create it in Riverflow
Animation 1: Rome Installation — 7 seconds
Use this structure to turn the strategy into a specific creative brief that keeps the product accurate and the scene useful.
- 1
Start frame
Use Frame 1, Rome Installation.
- 2
Model
Choose Seedance 2.5, set seven seconds, and keep the output vertical 3:4.
Create it in Riverflow
Animation 2: Hooded Beauty — 6 seconds
Use this structure to turn the strategy into a specific creative brief that keeps the product accurate and the scene useful.
- 1
Start frame
Use Frame 3, Hooded Beauty.
- 2
Model
Choose Seedance 2.5, set six seconds, and keep the output vertical 3:4.
Create it in Riverflow
Animation 3: MORA Swing — 7 seconds
Use this structure to turn the strategy into a specific creative brief that keeps the product accurate and the scene useful.
- 1
Start frame
Use Frame 5, MORA Swing.
- 2
Model
Choose Seedance 2.5, set seven seconds, and keep the output vertical 3:4.
Review checklist
Before you publish
Before generating the next frame
- The garment, label design, check pattern, fringe detail, and colourways are finished rather than placeholders.
- The product sheet shows the garment open, folded, the woven-label macro, and all colourways, and remains locked as the shared reference.
- The numbered composition, product-sheet, and model references follow the roles written in the full prompt.
- Every still uses Riverflow 2.5 and outputs exactly 3456 × 4320 pixels in vertical 4:5.
- Frame 2 receives the Nano Banana Pro retouch, and its composition and woven label are checked after the edit.
- Frames 1, 3, and 5 use their approved still as the start frame and the product sheet as the second reference in Seedance 2.5.
- Motion duration matches the prompt: seven seconds for Rome Installation and MORA Swing, six seconds for Hooded Beauty, all in vertical 3:4.
FAQ
Which model and format should I use for the six still images?+
Use Riverflow 2.5 and output every still at exactly 3456 × 4320 pixels in vertical 4:5. Frame 2 then receives a Nano Banana Pro retouch.
How do I create the product sheet?+
Create one Riverflow board showing the finished garment open, folded, the woven-label macro, and every colourway. The source guide requires this sheet but does not provide a product-sheet generation prompt.
Which references should I attach?+
Lock the product sheet as the shared garment reference and follow the numbered reference roles in each full image prompt. For motion, use the approved still as the first frame and the product sheet as the second reference.
Which model, duration, and format should I use for motion?+
Use Seedance 2.5 in vertical 3:4. Rome Installation and MORA Swing run for seven seconds; Hooded Beauty runs for six seconds.
Start creating
Get started with on-brand visuals
Turn guide ideas into product-accurate creative in Riverflow, using your brand, products, scenes, styles, and channel crops from the start.



