How to Create a Branded Stop-Motion Collage Video with Google Flow

A stop-motion collage video used to take days of cutting paper. AI can produce a video in minutes. Producing one that actually looks like your brand is the harder part.

For a recent Planet Media social video, we built four dense editorial collages from brand assets, animated them as tactile paper cutouts in Google Flow, and joined the resulting clips into one short video. The final piece feels handmade, energetic and unmistakably on-brand, not like a generic AI template.

In this tutorial, we are sharing the complete stop-motion collage video workflow, including reusable prompts that can be adapted to almost any company, product or campaign.

What Your Stop-Motion Collage Video Will Look Like

The finished video consists of four short scenes:

1. A rapid collage assembly that introduces the brand.
2. A second scene that shows the brand, audience or offer in action.
3. A clean branded message or campaign statement.
4. A final visual world that connects the brand to its products, sectors or desired outcome.

Each scene begins on textured paper, assembles rapidly, stays alive through small independent movements, and then clears away. The empty paper at the beginning and end of every clip creates a natural transition when the four scenes are joined.

The workflow is simple: plan the story, generate four still images, correct the brand assets, animate each still, combine the clips, then add audio and final text.

Tools You Need

You can complete the workflow with:

• An AI image generator that accepts reference images. Google Flow now includes image creation and editing in the same workspace, but you can also create the stills with another image model.
• Google Flow for image-to-video animation.
• Canva, Figma, Photoshop or another image editor for placing exact logos, packaging and website screenshots.
• CapCut, Premiere Pro, DaVinci Resolve or another video editor for joining the clips and adding music.
• A small brand kit with your logo, colors, fonts, screenshots, product photography and campaign message.

Google’s current Flow workflow can generate and edit images, use visual references as ingredients, animate frames and create video clips in different aspect ratios and lengths. Exact controls vary by model, subscription and region, so check the options visible in your project before generating. See Google’s guides for creating videos in Flow and creating or editing images in Flow.

Step 1: Define the Brand Inputs

Do not begin by asking an image model to “make a cool brand video.” First, give the project a small, fixed visual system. Complete this brand brief:

For this style, choose objects that can perform a visible action. A flower can unfold. A camera can flash. A record can rotate. A plant can sway. A paper arrow can draw itself. A product bottle can bounce into position. Aim for 10 to 15 independent objects per scene: enough to create constant movement without turning the composition into clutter.

Reusable Storyboard Planning Prompt

Paste the following prompt into your preferred writing assistant and replace the brackets:

I am creating a 20-24 second branded stop-motion paper-collage video for:

Brand: [BRAND NAME]
Industry: [INDUSTRY]
Audience: [AUDIENCE]
Offer: [PRODUCT, SERVICE OR CAMPAIGN]
Main message: [CAMPAIGN MESSAGE]
Brand personality: [3-5 ADJECTIVES]
Brand colors: [HEX COLORS]
Required assets: [LOGO, PRODUCT, WEBSITE, PACKAGING, PEOPLE]
Format: [4:5, 9:16 OR 16:9]

Create a four-scene storyboard. Each scene will become one six-second image-to-video clip.

For each scene provide:
1. Its communication purpose.
2. One clear focal point.
3. A list of 10-15 separate collage objects.
4. How each object can enter the frame.
5. One small continuous movement for each object.
6. How the scene can disassemble back to empty paper.

The four scenes should be:
1. Brand introduction.
2. Product, service or audience in action.
3. Simple campaign-message transition card.
4. Brand ecosystem, transformation or final outcome.

Avoid repeating the logo in the same scene. Do not rely on AI-generated text. Choose visually distinct objects that can be animated independently.

Step 2: Design the Four Static Collages

The most reliable AI video workflow starts with a strong final frame. Instead of asking the video model to invent the composition while it animates, create the exact still image you want first. That image becomes the visual target for the video, giving you far more control over layout, color, objects and brand consistency.

Choose the Format Before Generating

• Instagram feed: 1080 x 1350 px, 4:5.
• Instagram Reels, TikTok and YouTube Shorts: 1080 x 1920 px, 9:16.
• Website or YouTube landscape: 1920 x 1080 px, 16:9.

If your video model only offers 9:16 portrait output but you also need a 4:5 feed version, keep all essential elements inside a centered 4:5 safe area. You can then export the vertical version and create a separate feed crop without losing the focal point.

Build for Motion, Not Just Appearance

A beautiful static collage can still make a poor video. Give every major object room to move. Use separate cutout objects with visible edges, small gaps between important layers, one dominant focal point, shallow depth rather than realistic perspective, flat paper shadows, a consistent paper texture, objects at different scales for rhythm, and clear empty areas where elements can enter or leave.

Avoid several logos competing in one frame, tiny text that the video model will distort, large overlapping objects that cannot move independently, highly realistic depth of field, a single central object with nothing else to animate, and generating all four compositions as one grid. Generate and approve each image separately.

Master Static-Collage Image Prompt

Use this prompt as the foundation for every scene:

Create one premium vertical editorial paper collage for [BRAND NAME].

CAMPAIGN
Industry: [INDUSTRY]
Audience: [AUDIENCE]
Offer: [PRODUCT OR SERVICE]
Scene purpose: [BRAND INTRODUCTION / OFFER IN ACTION / MESSAGE CARD / FINAL OUTCOME]
Main visual idea: [DESCRIBE THE SCENE]

FORMAT
Aspect ratio: [4:5 OR 9:16]
Final size: [1080 x 1350 OR 1080 x 1920]
Create one finished image only, not a grid and not a storyboard sheet.

STYLE
Physical magazine cut-and-paste collage on off-white zine paper. Real photographic cutouts, torn paper, crisp white scissor-cut edges, xerox grain, halftone dots, risograph texture, matte printed paper, tape, marker accents, flat handmade contact shadows and shallow physical-paper depth.

BRAND PALETTE
[LIST THE BRAND COLORS AND HEX VALUES]

COMPOSITION
Use one clear focal point and 10-15 separate animation-ready objects.
Arrange the objects as distinct paper layers with enough space for each layer to slide, bounce, rotate, flutter or sway.
Keep the layout energetic but intentional, with a strong visual hierarchy and balanced negative space.

INCLUDE
[LIST THE EXACT OBJECTS FOR THIS SCENE]

BRAND-ASSET RULE
Use the attached product and website references as visual guides.
Do not invent a new logo, packaging label, interface or brand name.
Leave any area intended for the official logo as a clean [BRAND COLOR] printed-paper sign so the exact logo can be added manually afterward.
Leave text banners blank unless an exact, flattened text asset is supplied.

NEGATIVE PROMPT
Generic corporate stock image, glossy 3D, photorealistic desk scene, smooth vector illustration, repeated logo, invented logo, misspelled text, fake website, duplicated objects, extra people, crowded composition, heavy shadows, realistic depth of field, unrelated colors, watermark, image grid, storyboard sheet.

Four Adaptable Scene Directions

Add one of these blocks to the master prompt for each still.

Scene 1: Brand World

Create the brand introduction scene.

The focal point is [PRODUCT / LAPTOP WITH WEBSITE / FOUNDER / BRAND SYMBOL].
Surround it with objects that communicate the brand's personality, process and offer.

Suggested objects:
[EXACT PRODUCT], laptop or phone, website screenshot area, camera, notebook, sample materials, small plant, envelope, packaging, industry object, one natural element, one motion arrow, tape, paper scraps and halftone accents.

The composition should feel like the brand's complete world arriving on the page at once.

Scene 2: Brand in Action

Create the product, service or audience-in-action scene.

The focal point is [CUSTOMER / PRODUCT IN USE / CAMPAIGN RESULT].
Build a layered environment that shows where the brand creates value.

Suggested objects:
[CUSTOMER OR USER], product, phone, location fragments, transportation, buildings, dashboard or result graphic, signs, tools, plants, arrows, route lines, tape and paper textures.

The composition should communicate movement, usefulness and measurable progress.

Scene 3: Campaign Message

Create a cleaner transition-card scene.

Use fewer objects and more negative space than the other scenes.
Leave a central torn-paper banner for the exact campaign message:
[CAMPAIGN MESSAGE]

Include 5-8 supporting elements such as a bird, flower, product silhouette, tape, marker line, circle stamp, plus marks, halftone scraps and one large brand-color paper field.

Do not generate the campaign text. Keep the banner blank so the exact typography can be added and flattened manually.

Scene 4: Transformation or Outcome

Create the final brand-ecosystem or transformation scene.

Use a visual path, road, ribbon, river, sequence or layered landscape to guide the eye from the foreground toward the desired outcome.

The focal point is [FINAL RESULT OR CUSTOMER OUTCOME].
Include visual references to [PRODUCT CATEGORIES, SERVICES, MARKETS OR IMPACT AREAS].

The image should feel optimistic, complete and forward-moving, while remaining a tactile physical paper collage.

Step 3: Correct the Logo, Text and Screens Before Animation

This is the step most people skip, and the main reason branded AI videos fail. Image and video models may approximate a logo, change the typeface, misspell the name, duplicate the mark or turn a branded billboard into an empty white rectangle. A longer prompt does not always solve the problem.

The safest workflow is:

1. Generate the base collage.
2. Leave the logo area blank or remove the incorrect generated mark.
3. Open the image in Canva, Figma or Photoshop.
4. Place the exact transparent logo PNG into the composition.
5. Insert the real website screenshot, product label or campaign text.
6. Flatten the completed artwork into one PNG or high-quality JPG.
7. Use that corrected file as the animation reference.

Do not ask the video model to type the logo again. It should animate the complete printed object, not recreate the artwork inside it. One logo placement per scene is usually enough.

Logo and Screen Lock Prompt

Add this block to every animation prompt containing a logo, label, billboard or screen:

CRITICAL BRAND-ASSET LOCK

Use the exact logo, text, product label and screen pixels already embedded in the supplied reference image.

Treat each branded surface and its printed artwork as ONE SINGLE RIGID FLATTENED PAPER LAYER.

Do not generate, redraw, type, reconstruct, reinterpret or approximate anything inside that layer.

The complete branded object must enter with its final artwork already printed on it from its first visible frame.

The branded surface may move, bounce, rock, slide or jitter as one object, but its internal artwork must remain permanently attached, sharp and unchanged.

It must never become blank, white, empty, transparent, overexposed or replaced by generic content.

Do not animate anything inside the logo, website screenshot, product label or text banner.

Negative prompt:
alternative logo, fake logo, redrawn logo, generated typography, misspelled brand name, changed leaf or symbol, distorted wordmark, blank sign, white billboard, empty screen, fake website, changing interface, disappearing text, fading logo, duplicated logo, overexposed branded surface.

Step 4: Animate Each Image in Google Flow

Create a separate video generation for each of the four stills. In Google Flow:

1. Open an existing project or create a new one.
2. Add the completed still image to the project.
3. In the prompt controls, select Video.
4. For this build-and-disassemble effect, add the completed collage as a visual Ingredient or reference when that mode is available.
5. Choose a portrait aspect ratio for social media and a six-second generation length when available.
6. Paste the master animation prompt.
7. Generate at least two variations and keep the one that best preserves the composition.
8. Repeat for the other three stills.

Why use the final collage as a reference instead of simply making it the first frame? The animation needs to begin on empty paper, build toward the supplied composition, and clear away again. The still is the target composition, not necessarily the opening frame. If your Flow model only exposes Frames, use the collage as a start frame and create a simpler living-collage-plus-exit clip, or create the assembly and exit as two separate clips.

Master Image-to-Video Animation Prompt

Replace the bracketed fields and the object list for each scene:

Generate one image-to-video clip using the supplied [BRAND NAME] collage as the exact visual reference and final assembled composition.

OUTPUT
Duration: 6 seconds
Aspect ratio: [4:5 OR 9:16 PORTRAIT]
Style: physical stop-motion paper collage
Apparent animation rate: 8-12 fps

REFERENCE-IMAGE LOCK

Preserve every existing object, position, proportion, brand color and paper texture from the supplied reference.
Do not redesign, replace, duplicate or add objects.
The reference image is the exact visual target for the fully assembled scene.

The animation must:
1. Begin on nearly empty off-white textured paper.
2. Assemble rapidly into the supplied reference composition.
3. Reach at least 90% of the final composition before 1.2 seconds.
4. Keep at least 10-15 separate layers moving subtly during the middle section.
5. Disassemble and finish on empty paper.

STOP-MOTION STYLE

Use deliberately stepped low-framerate movement, approximately 8-12 fps.
Include small frame-to-frame paper jitter, handmade registration wobble, tiny jump cuts, uneven timing, paper-edge vibration, tape flutter, simple 2-3 frame cycles, shallow physical-paper parallax and flat contact shadows.
The result should feel like a real collage photographed frame by frame on a tabletop.
Do not use smooth digital interpolation.

TIMELINE

0.00-0.25 s: EMPTY PAPER. Begin on almost empty off-white paper with subtle grain and light xerox texture. Do not show blank signs, empty screens or unprinted product packages.

0.25-1.20 s: RAPID OVERLAPPING ASSEMBLY. Build almost the entire composition in one fast burst, with several objects entering at the same time. Do not create a slow object-by-object slideshow.
[OBJECT 1] [ENTRANCE ACTION].
[OBJECT 2] [ENTRANCE ACTION].
[CONTINUE FOR EVERY IMPORTANT OBJECT].
All logos, labels, website screenshots and text must already be printed on their complete objects when those objects enter. The final arrangement must match the supplied reference before 1.2 seconds.

1.20-5.10 s: LIVING COLLAGE. Keep the final composition stable and recognizable while the layers move independently.
[OBJECT 1] [SMALL REPEATING MOTION].
[OBJECT 2] [SMALL REPEATING MOTION].
[CONTINUE FOR AT LEAST 10-15 LAYERS].
Keep all movements small, tactile, stepped and loop-like. Do not distort the composition or regenerate the contents of any branded object.

5.10-6.00 s: PAPER EXIT. Disassemble the collage rapidly in staggered directions.
[DESCRIBE WHICH OBJECTS PEEL, FOLD, SLIDE, DROP OR FLY AWAY].
Every branded object must exit as one complete rigid printed-paper piece. End on completely empty off-white textured paper.

GLOBAL NEGATIVE PROMPT
Static still image, frozen middle section, slow slideshow, slow object-by-object entrance, simple zoom, simple pan, Ken Burns effect, smooth 60fps motion, fluid interpolation, polished corporate motion graphics, glossy 3D, cinematic photorealism, morphing objects, melting paper, elastic objects, warped cutouts, changing composition, duplicated objects, extra people, extra objects, invented logo, misspelled text, blank billboard, empty screen, fake interface, altered product packaging, heavy shadows, realistic depth of field, unrelated colors, watermark, cropped focal point.

Step 5: Give Every Object a Specific Movement

“Make the image move” is too vague. Assign a small physical action to each object.

The middle section is where many AI videos fail. If nothing moves after the composition assembles, the result feels like a still image with a transition. Explicitly request at least 10 independent micro-movements.

Step 6: Generate the Four Clips Separately

Do not ask Flow to create the entire 20 to 24 second piece in one generation. Working one scene at a time gives you better control over brand consistency, fewer deformed logos and screens, more precise motion, easier retries, cleaner editing, and the option to replace one weak shot without rebuilding everything.

Keep the same off-white paper background, collage style, animation rate, shadow treatment, color system, and general assembly and exit rhythm. Change only the objects and scene-specific movement.

Step 7: Join the Clips into One Video

Once you have four approved MP4 clips, combine them in CapCut or your preferred editor.

Simple CapCut Workflow

1. Create a new project.
2. Import the four clips.
3. Place them in story order: Scene 1, Scene 2, Scene 3, Scene 4.
4. Trim about 0.15 to 0.25 seconds of empty paper from the joins if the pause feels too long.
5. Do not add decorative transitions. The shared paper background already connects the clips.
6. Add one music track across the complete video.
7. Add any headline, caption or call to action in the editor, not inside the generated video.
8. Check the safe zones for the intended social platform.
9. Export as H.264 MP4 at 30 fps.

Recommended exports: 1080 x 1350 px for an Instagram feed, 1080 x 1920 px for Reels, TikTok and Shorts, and 1920 x 1080 px for website landscape. Using one music track in the editor usually produces a more coherent result than asking every six-second AI clip to create separate audio.

Troubleshooting a Stop-Motion Collage Video

The logo changes or uses the wrong typeface

Do not ask the model to spell or redraw the name. Place the official PNG in the final still, flatten the artwork and use the brand-asset lock prompt.

The billboard or screen turns white

Make the sign, its background and its artwork one rigid flattened layer before animation. Remove instructions for glow, flashes or changing screen content.

The logo appears twice

Choose one primary brand placement. Remove the weaker duplicate from the still image before animation.

The video feels static

List at least 10 to 15 independent movements. Give the paper edges, tape, marker lines and halftone texture subtle motion as well as the main objects.

The animation looks too smooth

Request an apparent 8 to 12 fps animation rate, small stepped jumps, uneven timing, registration wobble and 2 to 3 frame cycles. Add no smooth digital interpolation to the negative prompt.

The composition changes

Strengthen the reference-image lock. Reduce large camera movements and deep parallax. Ask for micro-movements rather than full character actions.

Objects enter too slowly

Use the phrase: rapid overlapping burst assembly; at least 90% of the composition is complete before 1.2 seconds.

Text becomes unreadable

Keep text out of the generative stage whenever possible. Add the final message in CapCut, Canva or Premiere. If text must exist inside the collage, flatten it as part of one rigid paper banner.

The four clips do not feel connected

Use the same paper background, palette, animation rate and exit style. End every clip on empty paper and trim the joins rather than adding unrelated transitions.

A More Intentional AI Workflow

Generative video requires experimentation, but more generations do not automatically create a better result. Plan the storyboard first. Test motion with faster or lower-cost settings when available. Correct the still image before regenerating the video. Upscale only approved clips. Reuse the same visual assets across the campaign.

The important idea is simple: use AI for movement and variation, but keep brand decisions under human control.

Final Quality-Control Checklist

• The official logo is exact and appears only where intended.
• Product packaging and website screens remain unchanged.
• No billboard or device becomes blank.
• All campaign text is spelled correctly.
• The four scenes use the same paper texture and color system.
• At least 10 objects move during each active scene.
• The motion feels stepped and handmade.
• No key object is cropped by the social safe zone.
• The video has one clear message and one final call to action.
• The exported file matches the intended platform dimensions.

Frequently Asked Questions

Turn Your Brand into a Repeatable Content System

A strong AI video is not just a one-off effect. The same approved images, objects, movement rules and prompt structure can become a repeatable visual system for product launches, campaigns, case studies and social content.

If your organization needs help combining brand strategy, creative direction and practical AI workflows, explore Planet Media’s AI and automation services for sustainable brands and marketing services for sustainable brands, or request a quote.

Kurt Whitt

Planet Media

Founder and CEO of Planet Media, a sustainability focused marketing agency. 25+ years helping purpose driven brands grow through strategy, storytelling and design.

More about the team