ScrollSlop — The Animate Guide

Why We Call It "Animate"

Every AI model has its own name for the same idea. If you've been researching video AI tools, you may have seen terms like:

Move (WAN 2.2)
Reference-to-Video (WAN 2.7)
Motion Control (Kling 3.0)
Motion Transfer / Image-to-Video (Seedance 2.0)

These are all the same core concept. We chose one word — Animate — so you don't have to look up what each company means.

Official Names Reference Table

If you've searched for one of these terms and landed here — you're in the right place.

What we call it	Model	What the model officially calls it
Animate	Seedance 2.0	Motion Transfer / Image-to-Video
Animate	Kling 3.0	Motion Control / Image-to-Video
Animate	WAN 2.2	Move / Image-to-Video (I2V)
Animate	WAN 2.7	Reference-to-Video (R2V)

What Animate Actually Does

When you hit Animate, here's what happens behind the scenes:

Your character image is loaded as the visual reference (who the avatar looks like)
Your driving video is analyzed frame by frame (what the avatar does)
The AI transfers the motion from your driving video onto your character
The result is your character, doing exactly what you did — your movements, your expressions, your energy

Same process, every model. The difference is in how each model handles detail, consistency, and the type of prompting it responds to best.

Available Models

Model	Best for	Character styles	Speed
Seedance 2.0	Natural motion, signing, fast drafts	Cartoon, illustrated, stylized	Fast
Kling 3.0	Cinematic quality, complex motion	Realistic, detailed, all styles	Medium
WAN 2.2	Open model, flexible motion	All styles	Medium
WAN 2.7	Character consistency across shots, series	All styles	Slower

Not sure which to pick? Start with Seedance 2.0 for most jobs. Use WAN 2.7 when series consistency matters most. Use Kling 3.0 for the highest visual quality.

Seedance 2.0

Also known as: Motion Transfer / Image-to-Video

How Seedance 2.0 Works

Seedance uses a reference-based system. Your character image is the @Image1 reference. Your prompt tells the model what to do with that reference.

The Golden Prompt Formula

[Character description] + [Action] + [Scene] + [Camera] + [Style] + [Consistency lock]

Example for ScrollSlop (content video)

A [character type] speaking directly to camera with expressive gestures, 
[background description], medium shot, natural lighting, smooth continuous motion, 
stable character, consistent face and outfit

Seedance-Specific Prompt Tips

Do use:

static camera or slow push-in — keeps focus on the character
natural lighting or soft studio lighting — clean output
smooth continuous motion — reduces jitter
stable face, consistent character — improves consistency between frames
natural anatomy, correct hands — helps the model prioritize hand accuracy
Specific scene descriptions: neutral grey background, indoor studio, outdoor park

Avoid:

Vague words like "beautiful" or "amazing" — use specific descriptions instead
Multiple competing instructions (don't say "close-up wide shot")
Asking for too much movement in one prompt

Seedance Prompt for ASL/Signing

Character signing in ASL, both hands clearly visible throughout, 
expressive face, neutral clean background, static camera, 
waist-up framing, natural lighting, smooth motion, 
stable anatomy, consistent character

Style Presets Explained

When you select a style preset, we add specific keywords to your prompt automatically:

Preset	What it adds
None	No additional style — pure motion transfer
Realistic	`photorealistic, natural skin tones, lifelike textures`
Stylized	`artistic rendering, enhanced contrast, stylized motion`
Cartoon	`flat colors, clean outlines, animated style, smooth motion`

For ASL/signing: "None" or "Cartoon" usually give the cleanest results. "Realistic" works best for photo-based character images.

Kling 3.0

Also known as: Motion Control / Image-to-Video

How Kling 3.0 Works

Kling 3.0 is designed to understand cinematic intent. Write your prompts like directions to a scene, not a list of objects. It responds to explicit motion instructions and character anchoring.

The Kling Prompt Formula

[Subject anchor] + [Action/motion description] + [Camera behavior] + [Scene/environment] + [Mood/style] + [Consistency instructions]

Example for ScrollSlop

A [character] facing camera and delivering dialogue with expressive gestures, 
camera tracks slightly forward then holds steady in medium close-up, 
[background], professional lighting, natural confident movement, 
consistent character design throughout

Kling-Specific Prompt Tips

Think in shots, not clips:

Kling handles multi-shot sequences well. Structure your prompt like a shot list:

First: character establishes in frame, medium shot. 
Then: character signs directly to camera, hands clearly visible. 
Finally: character pauses and looks forward.

Anchor your subject early:

Define your character at the very start of the prompt. Once established, Kling maintains consistency — face, clothing, style.

Explicit motion beats vague motion:

Instead of: the character moves

Write: the character raises both hands into signing position, gestures expressively, then lowers hands naturally

Camera language that works well:

medium shot, static camera — clean, stable
camera holds in medium close-up — focused on face and hands
slow push-in — slight zoom for emphasis
tracking shot, follows subject — for movement across frame

Kling for ASL/Signing

[Character description], both hands raised in natural ASL signing space, 
making deliberate clear handshapes, camera holds steady in medium shot, 
neutral indoor background, soft diffused lighting, 
smooth natural pacing between signs, 
same character maintained throughout, stable anatomy

Key for Kling: Kling is more responsive to detailed motion description than Seedance. Spend the extra words describing the hands specifically.

WAN 2.2

Also known as: Move / Image-to-Video (I2V)

What WAN 2.2 Does

WAN 2.2 is an open-source video model (originally called "Move" in many interfaces). It takes your character image and driving video and generates fluid motion. It's highly flexible and handles a wide range of character styles.

Prompt Formula

[Character description] + [Action] + [Scene] + [Camera] + [Motion quality]

Example

A [character type] signing clearly, both hands visible, 
neutral background, static camera, smooth natural motion, 
consistent character, stable face and hands

WAN 2.2 Prompt Tips

Straightforward descriptive prompts work well
Describe the motion you want explicitly: smooth, fluid, natural pacing
Works well with static camera for signing content
Add consistent character to reduce drift between frames
Less sensitive to cinematic language than Kling — focus on what you want to happen

When to Use WAN 2.2

When you want a reliable all-purpose result
Good for testing a prompt before running on other models
Strong with cartoon and illustrated characters

WAN 2.7

Also known as: Reference-to-Video (R2V)

What WAN 2.7 Does

WAN 2.7 introduces Reference-to-Video (R2V) — the model maintains your character's visual identity across every shot. This is the best model for series work where your avatar needs to look exactly the same video after video.

You provide a character reference image (or video), and WAN 2.7 uses it as a locked anchor. The character stays consistent even as motion, lighting, and scenes change.

Prompt Formula

[Reference anchor] + [Action/signing] + [Scene] + [Camera] + [Consistency lock]

Example

Reference character signing in ASL, both hands in natural signing space, 
clear expressive handshapes, clean neutral background, 
static camera at medium distance, warm natural lighting, 
smooth deliberate motion, consistent face and character throughout, 
identical appearance to reference

WAN 2.7 Prompt Tips

Always reference the character explicitly: consistent with reference image, identical character throughout
Great for multi-episode series — same prompt structure every time = reliable results
Describe what you want to stay the same as much as what you want to move
Works with up to 5 reference inputs — useful for series with established visual identity
Add stable anatomy and consistent outfit for tightest results

When to Use WAN 2.7

Building a series with the same character across many episodes
When character consistency is more important than speed
Long-form ASL content where your avatar must be recognizable every time

Resolution Guide

Setting	Output	Best for
480p	Fast, small file	Quick drafts, testing prompts
720p	Balanced quality	Most social media posts
1080p	Highest quality, longest time	Final versions, hero content

Recommendation: Use 480p to test your prompt and composition. Once happy with the result, re-run at 720p or 1080p for the final version. This saves credits.

Recording Your Driving Video

For Seedance 2.0 (Server 1)

Portrait orientation works best (vertical video)
Keep motion smooth and continuous — Seedance tracks fluid motion well
Avoid sudden stops or fast direction changes
Consistent lighting throughout the clip
Sign at a natural pace — not too fast

For Kling 3.0 (Server 2)

Both portrait and landscape work
Kling handles more complex motion — you can be slightly more expressive
Clear start and end position — Kling understands narrative structure
More expressive facial expressions transfer better with Kling
Works well with longer clips (up to 60 seconds)

For Both Models

Film in even lighting, no harsh shadows on hands
Hands fully in frame at all times
Solid colour top — no patterns
Simple or neutral background
If not wearing glasses in your character, remove glasses while recording
Hands: no rings, watches, or nail art that could confuse the AI
Stage your environment to match the final output you want

Common Mistakes and Fixes

❌ Hands look wrong or deformed

Character image has patterned or multi-coloured hands. Driving video doesn't keep hands in frame.

✅ Fix

Simplify character hands to a solid colour. Add correct hand anatomy, natural fingers, stable hands to prompt.

❌ Character face changes mid-video

Face drifts or morphs partway through the animation.

✅ Fix

Add same character throughout, consistent face, locked character design. For Kling: anchor character description at the start.

❌ Motion looks stiff or choppy

Animation doesn't flow naturally.

✅ Fix

Add smooth continuous motion, fluid movement, natural pacing. Check driving video — shaky = shaky output.

❌ Character doesn't look like my image

Output character looks different from your reference.

✅ Fix

Use a cleaner, simpler character image. Add faithful to reference character, match reference appearance.

❌ Signing doesn't transfer clearly

Signs are blurry or don't come through.

✅ Fix

Slow down signing in driving video. Add clear distinct handshapes, deliberate signing pace, both hands visible.

❌ Background bleeds into character

Character edges mix with the background.

✅ Fix

Use a character image with simple or no background. Add isolated character, clear subject separation.

Ready-to-Use Prompt Templates

Content Creator (Seedance 2.0)

A [character type] speaking to camera with confident expressive gestures, 
[background], medium close-up shot, professional lighting, 
natural smooth movement, stable character, consistent face and outfit

Cinematic Avatar (Kling 3.0)

A [character] delivering content with dynamic hand gestures and 
expressive facial animation, camera holds in medium shot then 
slow push-in for emphasis, [scene description], 
professional lighting, confident measured pacing, 
same character throughout, cinematic quality

Standard ASL Signing (Seedance 2.0)

A [cartoon / illustrated / realistic] character signing in ASL, 
both hands fully visible and expressive, neutral background, 
static camera, waist-up framing, natural even lighting, 
smooth continuous motion, stable face and anatomy, 
consistent character throughout

Standard Signing (WAN 2.2)

A [character type] signing in ASL, both hands fully visible, 
neutral background, static camera, smooth fluid motion, 
consistent character, stable face and anatomy

Series Episode — Consistent Character (WAN 2.7)

Reference character signing expressively in ASL, 
both hands in natural signing space, clear handshapes, 
clean background, static camera at medium distance, 
natural lighting, smooth motion, 
consistent with reference throughout, same face same outfit

⚡ Quick Reference Card

Goal	Model	Key prompt words
ASL signing — fast	Seedance 2.0	`static camera, both hands visible, natural signing space`
ASL signing — quality	Kling 3.0	`expressive hands, camera holds medium shot, cinematic`
Fingerspelling	Seedance 2.0 / WAN 2.2	`deliberate pace, each handshape distinct, hands centered`
Expressive content	Kling 3.0	`dynamic gestures, expressive face, slow push-in`
Series consistency	WAN 2.7	`consistent with reference, same face same outfit throughout`
All-purpose reliable	WAN 2.2	`smooth motion, natural lighting, consistent character`
Quick draft	Seedance 2.0 + 480p	`smooth motion, natural lighting`
Final quality	Kling 3.0 + 1080p	`cinematic quality, detailed character`