Seedance Before After logo

Seedance Before After

Community
rediumvex
seedance-before-after

Generate before-and-after transformation video prompts for Seedance 2.0 on Higgsfield. Use for transformation reveals, glow-ups, makeovers, renovation reveals, fitness transformations, design before-after, business growth visuals, or any content showing dramatic change. Triggers on before after, transformation, reveal, glow up, makeover, renovation, redesign, progress, results.

Overview

Publisherrediumvex
Repositoryai-video-generator-claude
Skill nameseedance-before-after
Stars
425
Forks
54
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by rediumvex on GitHub. Read the source before you install it.

Installation

Install the Seedance Before After AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/rediumvex/ai-video-generator-claude/tree/main/skills/07-before-after
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/rediumvex/ai-video-generator-claude.git /tmp/ai-video-generator-claude
mkdir -p .claude/skills
cp -r /tmp/ai-video-generator-claude/skills/07-before-after .claude/skills/seedance-before-after
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Seedance Before After in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Seedance Before After on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Seedance Before After is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Before-After — Transformation Reveal Video Prompts

Transformation content is the highest-engagement format on social media. This skill generates precise Seedance 2.0 prompts that capture the full emotional arc of change: tension, reveal, payoff.


Input Specs

Before generating prompts, establish:

  • Subject: What is transforming? (person, room, product, brand, body, space, logo, website)
  • From state: Describe the before — tone, condition, mood, colors, energy
  • To state: Describe the after — tone, condition, mood, colors, energy
  • Duration: 4s, 6s, or 8s clip
  • Aspect ratio: 9:16 (Reels/TikTok), 1:1 (feed), 16:9 (YouTube)
  • Transition style: hard cut, wipe, morph dissolve, object match, time-lapse compress
  • Platform destination: Instagram Reels, TikTok, YouTube Shorts, LinkedIn

Material References

  • Use @material[image1] for the "before" state photo
  • Use @material[image2] for the "after" state photo
  • Use @material[video1] for existing transformation footage
  • Use @material[audio1] for background music or sound effects

Platform Optimization

  • TikTok (9:16): Hook by 1.5s. Transition at 2-3s mark. Text overlays in center 80%. Bottom 20% obscured by UI.
  • Instagram Reels (9:16): Hook by 2s. Higher polish expected. Cover frame must work as grid image — pick the "after" hero moment.
  • YouTube Shorts (9:16): Hook by 2-3s. Can run longer (up to 60s). Description provides context.
  • LinkedIn (16:9 or 1:1): Professional tone. Slower transitions acceptable. Text overlays should include metric/result.

2-Second Hook Structures

1. The Flash Compare

Show both states in rapid alternation — before/after/before — within the first 1.5 seconds. The brain registers contrast before the viewer consciously decides to watch.

Prompt element: "rapid flash cut between [before state] and [after state], alternating twice in the first 1.5 seconds, creating a visual stutter effect that forces viewer attention"

2. The Slow Morph

Start on the "after" state. Hold for 1 second. Then slowly morph backward to the "before." Pause. Then morph forward to the "after" again. The reversal creates confusion and curiosity.

Prompt element: "begins on [after state], slow morph backward to [before state] over 1 second, pause, then forward morph to [after state], implying the transformation is reversible and therefore real"

3. The Destruction-to-Beauty

Open on maximum chaos, mess, or dysfunction. Hold for 0.5 seconds. Then a single triggering event (a hand gesture, a snap, a door opening) initiates the transformation sequence.

Prompt element: "opens on maximum [chaos/mess/dysfunction] state, 0.5 second hold, then [triggering gesture] initiates instant or rapid transformation to [beauty/order/completion]"

4. The Timeline Collapse

Show a compressed version of the entire journey — weeks or months — in the first 2 seconds via extreme time-lapse, then hold on the final result.

Prompt element: "hyper-compressed time-lapse of the full transformation journey in 2 seconds, frantic pace of change, then sudden freeze on the completed [after state] with a sharp focus pull"


Transition Techniques

The transition is the emotional peak of the video. Choose deliberately.

Hard Cut

An instantaneous frame-by-frame switch from before to after. Maximum impact. Best for dramatic, high-contrast transformations where the gap between states is enormous.

  • Use when: renovation reveals, body transformations, complete brand overhauls
  • Prompt element: "instantaneous hard cut from [before] to [after] at the exact midpoint of the clip, no transition frames, pure contrast"
  • Timing: place at 50-60% through the video to maximize buildup

Wipe

A directional reveal that sweeps across the frame, uncovering the after state beneath. The wipe direction carries meaning: left-to-right suggests progress, bottom-to-top suggests growth, diagonal suggests dynamism.

  • Use when: room makeovers, website redesigns, product packaging updates
  • Prompt element: "smooth [left-to-right / bottom-to-top / diagonal] wipe transition revealing [after state] beneath [before state], wipe speed synchronized with the beat of the background audio"
  • Timing: wipe should take 0.3–0.8 seconds depending on clip length

Morph Dissolve

The before state softly dissolves into the after, sharing screen real estate briefly. Creates an emotional, dreamy quality. Best when the transformation is about growth or healing rather than correction.

  • Use when: fitness transformations, personal growth, nature transformations, seasonal changes
  • Prompt element: "soft cross-dissolve morph from [before state] to [after state] over 0.8 seconds, both states share the frame at the midpoint creating an ethereal double-exposure moment"
  • Avoid: high-contrast color shifts (creates muddy midpoint)

Object Match

A specific object in the before state (a door handle, a can of paint, a barbell, a before photo) is used as a visual pivot. The camera or action traces that object, and the transition occurs mid-motion, hiding the cut in movement.

  • Use when: product reveals, personal trainer content, design work
  • Prompt element: "camera tracks [specific object] in continuous motion, transition hidden within the movement at the moment [object] is fully in frame, [before environment] seamlessly becomes [after environment] without breaking the motion"

Time-Lapse Compress

The transformation itself is shown at extreme speed — real footage or simulated — collapsing hours or months into seconds. Creates a visceral sense of effort and progress.

  • Use when: construction, fitness, creative projects, plant growth, cooking
  • Prompt element: "extreme time-lapse compression of [transformation process], 10x to 100x speed, motion blur on hands/tools/materials, smooth deceleration to real-time upon reaching the completed [after state]"

Visual Contrast Strategies

Contrast must be legible within the first second. Build it into every visual dimension.

Lighting Shift: Cold to Warm

Before state uses cool, flat, or harsh lighting (overcast daylight, fluorescent, blue-tinted). After state uses warm, golden, soft lighting (golden hour, warm tungsten, candlelight).

  • Before: "flat overhead fluorescent lighting, cool 5000K color temperature, harsh shadows, clinical"
  • After: "warm directional light at 3200K, soft golden fill, glowing skin tones, inviting atmosphere"

Color Shift: Gray to Saturated

Before state uses desaturated, muted, or monochromatic palette. After state explodes with rich, high-saturation color.

  • Before: "desaturated palette, near-monochrome, drained color, lifeless"
  • After: "fully saturated vivid colors, rich tones, every surface glowing with color energy"

Chaos to Order

Before state shows clutter, disorder, randomness, asymmetry. After state shows clarity, organization, symmetry, space.

  • Before: "visual chaos — objects at random angles, no clear sightlines, no negative space, cluttered surfaces"
  • After: "clean geometry, deliberate negative space, every object purposeful, sightlines open and clear"

Empty to Full

Before state is sparse, lonely, quiet, or barren. After state is populated, energized, vibrant.

  • Before: "empty room, bare walls, echoing silence, sparse and cold"
  • After: "space fully realized, personality present in every detail, warm and inhabited"

Camera Techniques for Reveals

Camera movement can amplify or undercut a transformation. Use these deliberately.

The Held Wide Shot: Lock the camera on a wide frame. Do not move. Let the transformation happen within the static frame. The stillness amplifies the drama of change. Best for room makeovers and exterior renovations.

"static wide shot, tripod locked, camera does not move throughout the entire clip, transformation occurs entirely within the fixed frame"

The Push-In Reveal: Start wide on the before state. Slowly push in toward the subject. The transition happens as the camera reaches maximum closeness. Pull back to reveal the after state already in place.

"slow push in from wide to close-up during [before state], transition at closest point, then slow pull back revealing [after state] in full"

The 360 Orbit: Camera orbits the subject continuously. The transition occurs as the camera passes behind or below the subject, hidden by the motion. The viewer sees before, the orbit continues, the viewer sees after.

"continuous 360-degree orbit around [subject], transition hidden as camera passes [behind/below] at the 180-degree mark, smooth uninterrupted motion throughout"

The Low Angle Rise: Camera starts at floor level looking up at the before state. Slowly rises. At the apex, or upon landing, the after state is revealed.

"camera begins at floor level, slow vertical rise, [before state] visible at start of rise, [after state] revealed fully at apex of camera movement"

The Split Screen Hold: Hold both before and after simultaneously in split screen for 1–2 seconds before merging into the full after state.

"split screen holds [before state] on left, [after state] on right, crisp vertical dividing line, both equally lit and framed, hold for 1.5 seconds before the dividing line sweeps off frame and [after state] fills the full canvas"


Sound Design: Contrast Audio

Audio carries 50% of the emotional weight in transformation content. Sync sound to visuals.

Chaotic to Clean

Before state: ambient noise, overlapping sounds, low-quality audio texture (hum, clutter, distraction). After state: clean, resonant, single-tone audio.

"audio transitions from ambient noise and clutter — indistinct overlapping sounds — to single clear resonant tone at the moment of visual transformation, silence for 0.5 seconds before music enters"

Silence to Music

Before state runs in near-silence or with only ambient sound. The transformation triggers music onset — ideally a beat drop or a melodic entry — that transforms the emotional register instantly.

"before state in ambient near-silence, transformation point triggers exact beat drop of [genre: cinematic, trap, acoustic, electronic], music at full volume from the first frame of the after state"

Low Energy to High Energy

Before state has slow tempo audio if any. After state has higher BPM, more energy, more presence.

"audio BPM doubles at the transition point, before state at 70 BPM ambient, after state at 140 BPM punchy, energy shift is immediate and total"

Sound Effect Punctuation

A single sound effect at the exact frame of transition acts as an audio exclamation point.

"single sharp [whoosh / click / snap / unlock sound] precisely synchronized with the visual transition frame, no ambient buildup, just the punctuation"


Complete Example Prompts

Example 1: Apartment Renovation Reveal (9:16, 6s)

Seedance 2.0 prompt — 6 seconds, 9:16, cinematic realism

Camera: static tripod-locked wide shot of a living room, camera does not move for the entire clip.

Before state (0:00–0:02): empty, rundown apartment living room. Peeling beige paint, cracked baseboard, single bare bulb hanging from ceiling, cold 5000K light, debris on floor, one broken chair leaning against far wall. Desaturated palette, near-monochrome. Heavy ambient silence with faint traffic noise.

Transition (0:02–0:03): a vertical wipe from left to right at a constant pace over exactly 1 second, the wipe line is a crisp edge with no glow or softness. As the wipe crosses the frame, it reveals the after state beneath.

After state (0:03–0:06): same room completely transformed. Warm 3200K indirect lighting, full furniture — linen sofa, wooden coffee table, art on walls, plants. Rich saturated palette, warm terracotta and deep green tones, golden hour light through new windows. At the exact frame the wipe completes, a single resonant interior design ambience — soft jazz piano, 90 BPM — begins at full presence. No fade in. Immediate music onset.

Hold on after state for 3 seconds. Shallow depth of field with gentle focus on sofa texture. Real-time, no motion.

Example 2: Fitness Transformation — Body Progress (9:16, 8s)

Seedance 2.0 prompt — 8 seconds, 9:16, clean studio aesthetic

Camera: slow push-in from mid-shot to close-up torso, starting wide, ending tight on subject.

Before state (0:00–0:03): person standing facing camera in a home setting. Slouched posture, casual oversized clothes, poor lighting from single overhead source. Flat body language, eyes slightly downcast. Muted color palette, shadows heavy, slightly underexposed. Ambient silence. Camera begins slow push-in.

Transition (0:03–0:04): camera reaches closest point to torso. Morph dissolve over 0.8 seconds — the before frame and after frame share the screen simultaneously at the midpoint, both at 50% opacity, creating a brief double-exposure effect before the after takes full presence.

After state (0:04–0:08): same person, same angle. Studio lighting at 45-degree split, warm key light from the right, cool fill from the left. Confident posture, fitted clothing, direct eye contact with lens. High muscle definition clearly lit. Rich skin tone saturation. Camera begins slow pull-back from close-up to wide over these 4 seconds. At the transition point, music begins — trap beat, 140 BPM, punchy kick on every 2 beats. Hold final wide shot for 1 second before cut.

Example 3: Brand Identity Redesign — Logo and Website (16:9, 4s)

Seedance 2.0 prompt — 4 seconds, 16:9, motion graphics hybrid

Camera: split screen for 2 seconds, then merge.

Before state left panel (0:00–0:02): left half of frame shows old brand. Cluttered logo with drop shadow and gradient, outdated 2010-era styling. Brand colors: muddy brown and dated gold. Website mockup visible beneath logo — busy layout, too many elements, no hierarchy. Cold light, desaturated. Low-quality render texture.

After state right panel (0:00–0:02): right half of frame shows new brand. Same split screen hold. Clean wordmark logo, confident sans-serif, minimal mark. Brand colors: sharp obsidian and warm cream. Website mockup below — generous whitespace, strong typographic hierarchy, one clear CTA visible. Warm directional light.

Transition (0:02–0:02.3): the vertical dividing line between panels begins to move right to left at high speed — a fast wipe that takes 0.3 seconds — sweeping the before state off screen, allowing the after state to expand and fill the full 16:9 frame.

After state full frame (0:02.3–0:04): new brand identity fills the entire screen. Website mockup animates in — hero text types itself character by character over 1 second. Logo pulses once gently with a single soft radial glow. A single clean UI click sound fires at 0:02.3 synchronized with the wipe completion. Then silence. Hold on full frame until cut at 4 seconds.

Example 4: Glow-Up Personal Style Reveal (9:16, 6s)

Seedance 2.0 prompt — 6 seconds, 9:16, fashion editorial quality

Camera: 360-degree orbit at chest height, continuous motion throughout the full 6 seconds.

Before state (0:00–0:02.5): person wearing everyday casual clothes — oversized hoodie, sweatpants, hair undone, minimal grooming. Natural indoor light, flat and directionless. Palette is muted neutrals. Camera is mid-orbit at the front of the subject, moving left. Expression is relaxed, unaware.

Transition (0:02.5): camera continues orbit, passing behind the subject. The transition is hidden entirely within the continuous rotation as the camera passes behind the body. This is the cut point — hidden in motion.

After state (0:02.5–0:06): as the camera re-emerges from behind the subject, the same person appears in a full styled look — tailored outfit, styled hair, intentional accessories, expressive makeup or grooming. Two directional studio lights at 45 degrees create strong modeling on the face and garment. Palette shifts to rich, high-contrast editorial colors. The moment the camera clears the back of the subject and reveals the styled look, a bass-heavy 140 BPM beat drops. The subject's expression is now confident and direct-to-lens. Camera continues orbit for another 180 degrees — a full front-to-back-to-front journey — completing the reveal at the 6-second mark.

Prompt Rules

Follow these rules for every Seedance 2.0 transformation prompt. Violating them produces weak or unusable output.

1. State the exact transition frame. Seedance needs to know exactly when the transition occurs. Always specify the timestamp: "transition at 0:03" or "transition occurs at the 50% mark." Vague instructions produce gradual fades when you want hard cuts.

2. Describe both states with equal specificity. A strong before description is as important as a strong after. If the before is vague, the contrast collapses and the reveal loses impact.

3. Specify lighting temperature numerically. "Warm" and "cool" are insufficient. Use Kelvin values: "2700K warm tungsten," "5000K flat daylight," "6500K cool overcast." Seedance responds to precise color temperature language.

4. Anchor the transition to camera movement or action. Free-floating transitions look unearned. Tie the transition to something: a hand gesture, a camera movement completing, an object passing the lens, a door opening. This grounds the edit in physical cause-and-effect.

5. Synchronize audio onset to the visual transition frame. State explicitly: "music begins at the exact frame of transition." Do not assume Seedance will infer this. If music should start on a beat drop, say so.

6. Specify aspect ratio and duration at the top of every prompt. Seedance 2.0 requires these as framing instructions. Without them, compositions may be incorrectly cropped for your platform.

7. Do not describe emotion — describe the conditions that produce emotion. Do not write "the viewer feels excited." Write "harsh fluorescent light becomes warm golden-hour glow." Emotion is the output. Light, color, posture, and sound are the inputs.

8. End on the after state with a hold. The final frame should be the best frame of the after state, held for at least 0.5 seconds. This is the shareable thumbnail frame and the dopamine delivery frame. Do not cut away before delivering it.

9. Use the transition to hide effort, not to show off technique. The best transformation transitions are invisible. The viewer should be surprised by the after state, not impressed by the wipe. When in doubt, use the hard cut or the motion-hidden object match.

10. One transformation per clip. Do not stack multiple transformation moments in a single clip unless you are explicitly using a Timeline Collapse structure. Multiple reveals dilute each other. One before, one after, one moment of maximum contrast.

Frequently asked questions

What does the Seedance Before After AI skill do?

Generate before-and-after transformation video prompts for Seedance 2.0 on Higgsfield. Use for transformation reveals, glow-ups, makeovers, renovation reveals, fitness transformations, design before-after, business growth visuals, or any content showing dramatic change. Triggers on before after, transformation, reveal, glow up, makeover, renovation, redesign, progress, results.

Why use Seedance Before After on TypingMind?

Because you install it once and use it with any model. Seedance Before After is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Seedance Before After in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/rediumvex/ai-video-generator-claude/tree/main/skills/07-before-after. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Seedance Before After?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Seedance Before After?

As many as you like. As long as a model supports skills, you can use Seedance Before After with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Seedance Before After AI skill free?

Yes. It is published on GitHub by rediumvex under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇