English | 中文 | 日本語

Awesome Seedance

Verified Seedance 2.5 / 2.0 prompt library. 463 cases checked against their original posts, 264 cross-model retests, 14 reusable templates and 25 installable AI-video Skills, drawn from goodcase.ai's 1229 verified AI cases across video, image, UI and copy. Synced daily, new cases land every day.

More verified AI cases with full prompts → GoodCase.ai

Contents

• Quick Links
• Install
• Why this list
• 🔁 Cross-model retests
• ⭐ Featured
• 🗂️ Category Overview
• 🧩 Prompt Templates
• 🔥 Top 30 by heat
• 🎬 All Prompts
• 🌐 Browse on goodcase.ai
• Statistics
• 🚀 How to use this repository
• How to Contribute
• 🙏 Acknowledgements
• Copyright & Takedown Notice
• Star History
• License & Reuse

Quick Links

Jump straight to the assets. The Contents list below is the section map of this page.

• Gallery index - All 463 cases with full prompts, every page in one place.
• Seedance 2.5 - 56 cases.
• Seedance 2.0, part 1/2 - Cases 1–95.
• Seedance 2.0, part 2/2 - Cases 96–130.
• Seedance (version unspecified), part 1/4 - Cases 1–91.
• Seedance (version unspecified), part 2/4 - Cases 92–193.
• Seedance (version unspecified), part 3/4 - Cases 194–265.
• Seedance (version unspecified), part 4/4 - Cases 266–277.
• Prompt templates - 14 reusable structures in 6 categories.
• Agent Skill - Install with npx seedance-prompt-library install into Claude Code / Codex.
• More AI-video Skills on goodcase.ai - 24 installable Skills grown out of the video cases.
• Live site on goodcase.ai - Search, heat leaderboard, stability ranking, retest logs.

Install

npx seedance-prompt-library install

Installs the seedance-prompt-library Agent Skill into Claude Code and Codex, so your agent can pull structured prompt templates and write Seedance prompts directly in your editor. Prefer the skills CLI? npx skills add LearnPrompt/awesome-seedance --skill seedance-prompt-library installs the same skill.

Want your agent to query the whole goodcase.ai library (all models, live data, retest baselines) instead of just the Seedance templates? Install the goodcase Skill from the sister repo LearnPrompt/goodcase-lite: npx skills add LearnPrompt/goodcase-lite --skill goodcase.

Why this list

Human-verified against the source. Every prompt here was checked against the creator's original post. Prompts reverse-engineered from the output video only, with no source and no submission, are rejected outright, per goodcase.ai's collection standards (in force since 2026-08-05).

Re-run on a second model. Most cases have been re-generated on another video model, with the verdict, score and output published. See Cross-model retests.

Full provenance on every entry. Author, original post link, publish date, and a heat score, a relative percentile among published cases on the same platform. If it didn't rank, it isn't here.

Ships as an installable Agent Skill. npx seedance-prompt-library install drops a template library straight into Claude Code / Codex so your agent writes Seedance prompts from proven structures, not guesses.

🔁 Cross-model retests

As far as we know, this is the first public prompt library that re-runs its video prompts on a second model at scale and publishes the result either way. 254 of the cases here have been re-run (264 runs so far), each with a verdict, a judge score and the generated output. A prompt that only ever worked once, for its author, on one model, is a screenshot; a prompt that survives a re-run is a method.

Model	Runs	Reproduction rate
MiniMax H3 Max 768p	253	73%
MiniMax H3 768p	11	82%

Verdicts across all runs: ✅ 193 reproduced · ⚠️ 68 degraded · ❌ 3 failed. Runs without a final score show as score n/a. Per-case verdicts, scores and output videos are on each case's goodcase.ai page; the model labels and batch dates are explained under Statistics.

Same prompt, second model. Three examples, including one that did not hold up:

Case	Original (Seedance)	Retest (second model)	Verdict
Seoul Summer Night Vlog 2.5 · heat 99		MiniMax H3 Max 768p ▶ output video	✅ reproduced (score 75.7)
Mini DV Coffee ASMR Vlog 2.5 · heat 99		MiniMax H3 Max 768p ▶ output video	✅ reproduced (score 77.3)
Frozen Time and Rewind in a 1950s Diner 2.5 · heat 100		MiniMax H3 Max 768p ▶ output video	⚠️ degraded (score 63.9)

Every run costs real inference money: over US$300 across 264 runs so far, at list price with no discounts, which is what anyone else would pay to reproduce them. We publish the result whether or not it flatters the prompt. Want Kling, Veo, Hailuo or the Seedance 2.5 API added to the retest matrix? Sponsor a batch → or write to [email protected].

⭐ Featured

Top 6 prompts by heat score, across all Seedance versions. Long prompts are collapsed; click to expand.

Frozen Time and Rewind in a 1950s Diner

A 30-second Seedance 2.5 script that locks a diner spill at its peak, orbits the suspended coffee ribbons for a full circle, rewinds every person and droplet back to the exact starting positions, then erases the accident with a single casual wave.

Author: @techhalla | Source: Original | Published: 2026-08-01 | Heat: 100

Stability: 64/100

Retest: MiniMax H3 Max 768p · 2026-09-07 · ⚠️ degraded (score 63.9) · output

Full prompt (13 lines, click to expand)

Photorealistic cinematic 1950s American diner, chrome stools, red vinyl, neon glow and checkerboard floor, shot with modern lived-in realism and soft natural window light. Subtle handheld texture, warm practicals, rich period detail, heavy film grain.

0-4s: [Medium Wide] A striking young woman in her early 20s sits alone at the counter, calm and slightly amused, slowly sipping a tall thick milkshake through a straw. Behind her a young waitress in classic uniform approaches with a tray of eggs and bacon in one hand and a full glass coffee pot in the other. An older lady starts rising from a nearby booth.

4-8s: [Dynamic Tracking] The older lady collides hard into the waitress. Tray, plate, eggs, bacon and coffee pot explode upward in chaotic slow motion. Coffee erupts into long liquid ribbons and perfect suspended droplets. Camera immediately begins a smooth continuous orbit around the impact. Time locks completely at the peak of the spill. Every face freezes in pure shock. Only the girl at the counter keeps moving, completely unfazed.

8-17s: [Slow 360° Orbital] Camera glides in a full elegant orbit through the frozen diner. Coffee hangs in mid-air as glassy ribbons and spheres with perfect volume and surface tension. Bacon strips, eggs and the spinning tray float weightlessly. Patrons and waitress remain locked in startled expressions. The girl at the counter takes one slow, deliberate sip, eyes half-lidded, almost bored, while the entire frozen world (except her) begins an elegant reverse: every droplet, every piece of food and every person rewinds smoothly back to the exact starting positions.

17-24s: [Medium Shot] Rewind lands perfectly. Waitress stands balanced again with tray and coffee pot. The girl lifts her eyes, raises two fingers in a small casual gesture and softly calls the waitress by name. The waitress turns toward her just before the older lady begins to stand, completely avoiding the collision. A tiny private smile crosses the girl’s face.

24-30s: [Extreme Close-Up] Hard cut to her face as she takes one last slow sip. Soft knowing smile, eyes almost closed in quiet satisfaction, like she has done this a hundred times. Shallow depth of field, creamy bokeh of the neon diner behind her.

Photorealistic, ultra-detailed fluid physics, perfect motion blur only on moving elements, stable characters, cinematic lighting, heavy natural film grain, no artifacts, movie-level temporal coherence, high rewatch value.

🔍 View on goodcase.ai (retest log / stability score) →

Seoul Summer Night Vlog

A thorough scripted prompt for a nostalgic home-video-style clip of a summer night in Seoul.

Author: @AIwithkhan | Source: Original | Published: 2026-08-27 | Heat: 99

Stability: 76/100

Retest: MiniMax H3 Max 768p · 2026-09-07 · ✅ reproduced (score 75.7) · output

Create a 30-second, 1080p ultra-realistic personal home-video showing an ordinary summer evening in the life of a young Korean woman. No reference image. MAIN SUBJECT Young Korean woman in her early 20s, naturally pretty, realistic skin texture, minimal makeup, relaxed and approachable personality. Long black hair loosely tied into a messy side ponytail with a few loose strands around her face. Wearing a fitted pastel-blue short top, loose cream pajama-style pants, black sneakers and a simple silver necklace. Maintain the same face, hairstyle, clothing, body proportions and overall appearance throughout the entire video. SETTING A quiet older Seoul residential neighborhood during a warm summer evening. Narrow concrete lanes, small houses, potted plants, bicycles, old walls, utility poles, overhead wires, a tiny neighborhood bakery, a public water tap and large leafy trees casting shadows across the street. Everything should feel lived-in, ordinary and peaceful. CAMERA / VISUAL AESTHETIC Raw personal footage casually recorded by a friend on an early-2000s consumer DV camcorder. Strong handheld shake, imperfect framing, autofocus hunting, exposure shifts, occasional motion blur, faded colors, soft digital detail, mild noise, accidental zooms and natural camera imperfections. — OUTSIDE THE HOUSE She steps outside carrying a small reusable shopping bag. She locks the door, adjusts her messy ponytail and looks toward the camera with a relaxed smile. — BAKERY STOP She reaches a tiny neighborhood bakery and buys a warm pastry. She steps outside, takes her first bite and immediately smiles because it tastes good. She notices her friend filming and holds the pastry toward the camera playfully before taking another bite. — WALKING HOME She continues down the narrow street while eating. A neighborhood cat follows her for a few steps. She notices it, crouches down and gently pets it. — SMALL ACCIDENT She walks beneath a large tree when a few leaves fall onto her hair. She stops, looks confused, then realizes what happened and laughs. — QUIET MOMENT She reaches a low concrete wall beside the street and sits down for a moment. — FINAL MOMENT She stands up and continues walking home.

🔍 View on goodcase.ai (retest log / stability score) →

Two-Idol K-pop MV, Shot by Shot

A 30-second K-pop music video cut into a dozen two-to-four-second beats, each pinning down camera distance, backdrop and choreography, with pink hair and black hair used as identity anchors so the two performers stay consistent across every cut.

Author: @Just_sharon7 | Source: Original | Published: 2026-08-01 | Heat: 99

Stability: 74/100

Retest: MiniMax H3 Max 768p · 2026-09-07 · ⚠️ degraded (score 74.2) · output

30-second ultra-realistic K-pop MV featuring two young East Asian women with flawless synchronization, cinematic lighting, glossy skin, realistic hair and fabric physics, natural body motion, and 4K live-action quality. Vibrant hot pink, electric blue, and silver color palette. 0–2s: Wide shot in a bright circular pink studio with reflective floor. Pink-haired woman (left) and black-haired woman (right) perform energetic opening pose and synchronized dance. 2–4s: Medium close-up of the black-haired woman on a blue spotlight stage, confidently pointing at the camera. 4–6s: Pink-haired woman dances before shimmering blue-silver tinsel curtains, dramatic hair flip and fluid arm movements. 6–8s: Back to the pink studio. Both perform synchronized choreography with sharp arm waves, hip sways, and strong formations. 8–10s: Extreme close-up of both faces against a blue background, glossy makeup, subtle smiles, and direct eye contact. 10–14s: Solo shots at the tinsel backdrop. Pink-haired woman mouths lyrics and gestures confidently, followed by the black-haired woman with relaxed jacket styling. 14–18s: Pink studio. Coordinated jacket choreography, hair flips, powerful synchronized dance, ending hands-on-hips. 18–22s: Glamour close-ups. Black-haired woman under glittering bokeh lights, then pink-haired woman with wind-blown hair against a soft pink background. 22–24s: Blue spotlight stage. Mirrored black-haired performer effect with synchronized spins and flowing hair. 24–26s: Both walk confidently toward the camera in front of shimmering tinsel curtains, reflections visible on the floor. 26–29s: Final synchronized dance and ending pose in the pink circular studio, standing together and looking into the camera. Style: Hyper-realistic live action, Seedance 2.5-quality motion realism, perfect lip sync, natural weight shifts, flowing hair, realistic fabric simulation, polished K-pop music video cinematography.

🔍 View on goodcase.ai (retest log / stability score) →

Mini DV Coffee ASMR Vlog

A cozy espresso-making vlog that manufactures authenticity out of consumer camcorder flaws — hand shake, focus hunting, auto-exposure shifts and tape grain — paired with an ASMR sound list, a three-second-per-beat storyboard and an explicit rule against looking commercial or AI-generated.

Author: @Strength04_X | Source: Original | Published: 2026-07-31 | Heat: 99

Stability: 77/100

Retest: MiniMax H3 Max 768p · 2026-09-07 · ✅ reproduced (score 77.3) · output

Full prompt (17 lines, click to expand)

CAMERA / LOOK: Handheld mini DV camcorder footage filmed by the subject herself. Slight hand shake, occasional focus hunting, imperfect framing, natural zoom adjustments, soft tape-like image quality, subtle grain, realistic auto-exposure shifts from bright kitchen morning light. Natural skin tones, mild motion blur, authentic consumer camcorder aesthetic.
STYLE: Cozy coffee-prep vlog with gentle ASMR elements. Relaxed pacing, minimal dialogue, candid moments. Focus on satisfying sounds: bean grinder whirring, portafilter tamping, steam wand hissing, cup clinking, milk frothing.
SUBJECT: Young man in his mid-20s, plain t-shirt, hair slightly tousled, minimal accessories. Calm, focused energy while making his morning coffee.
SETTING: Small kitchen counter with an espresso machine on a bright morning. Natural daylight, coffee beans and a mug nearby, quiet atmosphere.
STORYBOARD:
→ (3s, propped medium shot) Places camera on the counter, switches on the machine. "Morning coffee, the proper way."
→ (3s, overhead shot) Grinds fresh coffee beans, fine grounds falling into the portafilter.
→ (3s, close-up) Tamps the grounds down firmly and evenly.
→ (3s, handheld shot) Locks the portafilter into the machine. "Here we go."
→ (3s, detail shot) Espresso streams slowly into a small cup. No dialogue.
→ (3s, medium shot) Pours cold milk into a small steel pitcher. "Time for the milk."
→ (3s, macro shot) Steam wand hissing as it froths the milk.
→ (3s, propped shot) Pours frothed milk carefully over the espresso, forming light layers.
→ (3s, warm ending shot) Holds the finished cup, takes a small sip, satisfied smile. "That's exactly what I needed."
→ (3s, final shot) Reaches toward camera, still holding the cup. "See you later." Hand covers lens as recording ends.
AUDIO NOTES: Natural kitchen ambience — grinder whirring, tamping, steam hissing, milk pouring should be clearly audible. Dialogue quiet and casual.
REALISM NOTES: Authentic body language, natural blinking, genuine focused smiles, occasional careful pauses while pouring, imperfect framing, focus breathing, bright morning light shifts. Should resemble a genuine personal coffee vlog on a consumer camcorder, not a commercial or AI-generated production.

🔍 View on goodcase.ai (retest log / stability score) →

Thirty-Second Emotional Live-Action Discovery Scene

Pure emotions 😭 Can Hollywood match it ? Seedance 2.5 Prompt : Create a 30-second, 1080p ultra-realistic emotional live-action scene about a young woman discov…

Author: @AIwithSynthia | Source: Original | Published: 2026-09-06 | Heat: 95

Full prompt (54 lines, click to expand)

Create a 30-second, 1080p ultra-realistic emotional live-action scene about a young woman discovering her boyfriend kissing another woman inside a quiet café. Focus entirely on authentic human emotion, natural crying, subtle facial expressions and believable body language. No melodrama or exaggerated acting.
MAIN CHARACTER
Young Korean woman in her early 20s, naturally beautiful, realistic skin texture, minimal makeup, long black hair loosely styled, wearing a fitted pastel-pink top, casual denim shorts and simple sneakers.
Maintain the exact same face, hairstyle, outfit, body proportions and appearance throughout the entire scene.
LOCATION
A cozy modern café on a rainy evening. Warm interior lighting, large windows covered with rain droplets, wooden tables, soft background chatter, cups and plates, a few customers quietly talking.
SCENE
She enters the café carrying a small handbag, casually looking around for her boyfriend.
She suddenly notices him sitting at a table near the window.
He is kissing another young woman.
She immediately stops walking.
Her expression changes from confusion to disbelief.
She slowly lowers her handbag.
She stares at them silently for a few seconds, struggling to process what she is seeing.
Her eyes begin to fill with tears.
Show realistic tears naturally forming along her lower eyelids, then slowly rolling down her cheeks. No exaggerated crying.
She takes a shaky breath.
She walks toward their table.
Her boyfriend notices her.
His expression immediately changes to shock.
The other woman turns around and becomes uncomfortable.
The girlfriend looks directly at him with trembling eyes and quietly says:
“How could you do this to me?”
Her voice cracks naturally.
She wipes one tear from her cheek, but more tears continue falling.
Her boyfriend stands up and tries to explain.
She shakes her head and takes a small step backward.
She says through tears:
“I trusted you.”
She looks at him for one final moment.
Her expression shows heartbreak, disappointment and disbelief rather than anger.
She turns around and walks toward the café door while quietly crying.
Her boyfriend takes one step after her but stops.
She reaches the door, pauses briefly, wipes her face and leaves.
The camera remains inside the café as the door closes behind her.
CAMERA / VISUAL STYLE
Handheld cinematic camera with subtle natural movement. Start with a medium shot following her into the café, then slowly move closer when she notices the couple.
Use realistic close-ups of her eyes, trembling lips, tears and subtle facial reactions.
Natural shallow depth of field, realistic skin pores, authentic reflections in the café windows, soft rainy evening light, natural motion blur, 24fps, photorealistic live-action.
Do not make her crying look beautiful or glamorous. It should look spontaneous, uncomfortable and emotionally real.
AUDIO
Rain against the café windows, quiet conversations, cups being placed on tables, distant café ambience, footsteps and subtle breathing.
No background music.
No narration.
Dialogue should sound natural with slight pauses, trembling voice and emotional breathing.
EMOTIONAL DIRECTION
The emotion should build gradually:
confusion → realization → shock → disbelief → tears → confrontation → heartbreak → quiet acceptance.
Avoid screaming, physical aggression or exaggerated dramatic gestures.
CONTINUITY / REALISM
Keep tears physically consistent once they appear. Tears should remain visible on her cheeks and naturally accumulate as she continues crying.
Maintain consistent clothing, hairstyle, jewelry and facial identity.
No sudden emotional changes, no disappearing tears, no duplicated people, no changing café layout, no unnatural body movements.
NEGATIVE: plastic skin, fake tears, exaggerated crying, cartoon expressions, melodramatic acting, screaming, physical violence, distorted faces, extra fingers, identity drift, outfit changes, CGI

🔍 View on goodcase.ai (retest log / stability score) →

Jeweled Scorpion Transforms into a High Heel

A magical transformation where dark luxury meets elegance The scorpion turns into a stunning jeweled stiletto creating a

Author: @juliaevee | Source: Original | Published: 2026-08-27 | Heat: 95

A close-up vertical shot in 4K resolution depicting a magical transformation. A woman's bare foot steps down onto a large, ornate scorpion made of polished black obsidian, dark red rubies, and gold metal accents resting on a polished marble floor. As her foot touches it, the metallic ruby scorpion seamlessly morphs, wraps around her foot, and transforms into a luxury high-heeled shoe and anklet. The scorpion's tail curves around her ankle into a gem-studded strap, its claws form the pointed toe box, and its body shapes into a sleek, jeweled stiletto heel. Cinematic lighting, photorealistic, 3D render, luxury fashion aesthetic, rich depth of field, bright sunlight filtering through a grand greenhouse palace in the background, 8k ultra-detailed.

🔍 View on goodcase.ai (retest log / stability score) →

🗂️ Category Overview

Start from the look you want, then open that category's templates to turn it into a reusable structure. Each tile links to the templates below and to the verified cases behind them.

🏗️ Timeline & shot scripts 1 template · 4 verified cases Second-by-second timeline script View templates · Top case	🏗️ Reference & identity lock 1 template · 4 verified cases Reference image identity lock View templates · Top case	📱 Handheld UGC vlog 1 template · 4 verified cases Handheld UGC vlog View templates · Top case	📱 First-person POV take 1 template · 4 verified cases First-person continuous take View templates · Top case
🛍️ UGC creator review 1 template · 3 verified cases UGC creator review with spoken lines View templates · Top case	🛍️ Product commercial 1 template · 4 verified cases Cinematic product commercial shot list View templates · Top case	🛍️ Process & transformation 1 template · 4 verified cases Process and transformation montage View templates · Top case	🎭 Dialogue & lip-sync 1 template · 4 verified cases Dialogue and performance beats View templates · Top case
🎭 Cinematic narrative short 1 template · 4 verified cases Cinematic narrative short View templates · Top case	🎨 Anime & stop motion 2 templates · 8 verified cases Anime and stylized style lock · Stop motion and stepped cadence View templates · Top case	💥 Combat & physics set pieces 2 templates · 7 verified cases Combat choreography · Time freeze and rewind set piece View templates · Top case	💥 Beat-synced music video 1 template · 4 verified cases Beat-synced music video View templates · Top case

🧩 Prompt Templates

Reusable prompt structures distilled from the highest-performing cases, grouped by category. Each row is one template; the full structure, guidance and pitfalls for every template live in the Skill reference and ship with npx seedance-prompt-library install.

🏗️ Structural foundations (2 templates)

Cross-cutting skeletons that almost every other template builds on: how to slice a clip into timed beats, and how to pin an identity across those beats.

Template	Use when	Examples
Second-by-second timeline script Split the clip into contiguous timed segments, each carrying one shot type, one main action and its own sound line. The single most load-bearing structure in the corpus.	Any clip longer than about 8 seconds, or any clip where a specific thing must happen at a specific moment. 63 of 207 cases (30%) use timed segments, and the share rises to 45% among Seedance 2.5 cases.	#1 #2 #3 #4
Reference image identity lock Name every reference with a stable token, enumerate what to inherit from it, and separately enumerate what must not be inherited. The inherit-nothing-else clause is what separates working locks from broken ones.	Any clip where a face, an outfit, a product or a UI layout must survive across shots. Applies to Seedance 2.0 and 2.5 alike; 2.5 additionally accepts audio and video references under the same token scheme.	#1 #2 #3 #4

📱 Realism and UGC (2 templates)

Templates that buy believability by describing camera flaws, body wear and consumer-grade optics instead of asking for quality.

Template	Use when	Examples
Handheld UGC vlog Buy believability with camera defects. Name a specific consumer camera era, list its flaws as requirements, and switch cinematic polish off by hand.	Personal-feeling footage: daily life, travel, gym, cooking, get-ready-with-me. Use it whenever the goal is looks like someone actually filmed this rather than looks expensive.	#1 #2 #3 #4
First-person continuous take Bodycam, GoPro, FPV and handlebar POV. The camera is mounted on a body, so its motion has to be derived from that body, and every cut has to be declared by hand.	Immersive footage where the viewer is the operator: tactical entry, action sports, cooking from the cook's eyes, drone flight. 32 of 207 cases sit here.	#1 #2 #3 #4

🛍️ Commercial and product (3 templates)

Ad-shaped structures where a product has to survive macro shots, hand contact and a hero frame without deforming.

Template	Use when	Examples
UGC creator review with spoken lines A creator unboxes, handles and endorses a product on camera. Two independent locks are needed — one on the person, one on the product — and the spoken lines are welded into the actions.	Affiliate-style product videos, unboxings and creator reviews where the product must stay recognisable while being picked up, rotated and worn.	#1 #2 #3
Cinematic product commercial shot list A polished 8 to 20 second ad: a stated commercial aesthetic up front, a numbered or timed shot breakdown in the middle, a hero frame at the end, and a keyword tail.	Beauty, beverage, jewellery, automotive and fragrance spots where the look has to read as paid production rather than as a creator video.	#1 #2 #3 #4
Process and transformation montage Cooking steps, renovation timelapse, blueprint-to-house, miniature city assembly. The craft here is declaring what must not change, then ordering the change spatially.	Any clip whose subject is a process rather than a person: recipes, builds, assemblies, before-and-after reveals. 35 cases in the corpus.	#1 #2 #3 #4

🎭 Narrative and performance (2 templates)

Templates where the payload is a story beat or a line of dialogue rather than a look.

Template	Use when	Examples
Dialogue and performance beats Declare the spoken language, tag the speaker, write the reaction as a causal chain rather than a list of expressions, and close each beat with an explicit end state.	Whenever a line has to be heard rather than implied. 78 of 207 cases carry quoted dialogue inline (38%), and 16 explicitly manage lip sync. Seedance 2.5 additionally supports driving lip sync from an uploaded audio track.	#1 #2 #3 #4
Cinematic narrative short Multi-act storytelling in 15 to 60 seconds. Titled acts, a character card ahead of the acts, and a reveal written as a concrete image rather than as a promise of surprise.	Trailers, mini-dramas, disaster set pieces, sci-fi mysteries and romance shorts — anything where the viewer should follow a plot rather than admire a look.	#1 #2 #3 #4

🎨 Stylized animation (2 templates)

Non-photoreal looks that collapse into generic CG unless the style is specified as measurable parameters plus an exclusion list.

Template	Use when	Examples
Anime and stylized style lock Specify the drawing style as measurable parameters, then attach an exclusion list of the neighbouring styles it must not fall into. Without the exclusion list, anime collapses into a generic 3D face.	Cel-look action, Ghibli-flavoured slice of life, 3D toon RPG battles, 2D hand-drawn cooking. Roughly a quarter of the corpus is stylized animation of some kind.	#1 #2 #3 #4
Stop motion and stepped cadence Stop motion is a timing spec before it is a look. Pin the frame rate and the hold count, name the craft material, and ban the three things that silently smooth it away.	Claymation, paper-cut, moving-oil-painting, collage and tabletop object animation. 24 cases in the corpus sit in this family.	#1 #2 #3 #4

💥 Action, dance and effects (3 templates)

Templates driven by body mechanics, beat placement or a physics set-piece rather than by scene description.

Template	Use when	Examples
Combat choreography Fights read as real when the prompt specifies biomechanics, contact points and an attack chain. Adjectives like epic produce two people swinging at air.	Martial arts, swordplay, street fights, superhero traversal and stunt sequences. 40 cases in the corpus, split fairly evenly between live-action and anime treatments.	#1 #2 #3 #4
Beat-synced music video Derive beat anchors from BPM, pin every cut, hair flip and formation change to a real downbeat, and constrain the backup dancers so they never steal the visual centre.	K-pop MVs, dance covers, beat-cut fitness edits and club performance clips. Use the audio-anchored variant only on Seedance 2.5, which accepts an audio track as an input modality.	#1 #2 #3 #4
Time freeze and rewind set piece A five-beat skeleton — normal, collision, freeze at the peak, orbit, precise rewind — with exactly one character exempt from the freeze. The corpus contains the same author reusing this skeleton with a different physical material, which is direct evidence that it transfers.	Short high-engagement set pieces built on a physics spectacle rather than a plot. The diner version is the single highest-engagement case in the whole corpus.	#1 #2 #3

🔥 Top 30 by heat

The 30 hottest cases across all versions (ranks 1–6 are also shown in full under ⭐ Featured). prompt opens the full entry in the gallery, source opens the creator's original post.

#	Preview	Case	Version	Heat	Retest	Links
1		Frozen Time and Rewind in a 1950s Diner	2.5	100	⚠️ degraded (score 63.9)	prompt · source
2		Seoul Summer Night Vlog	2.5	99	✅ reproduced (score 75.7)	prompt · source
3		Two-Idol K-pop MV, Shot by Shot	2.5	99	⚠️ degraded (score 74.2)	prompt · source
4		Mini DV Coffee ASMR Vlog	2.5	99	✅ reproduced (score 77.3)	prompt · source
5		Thirty-Second Emotional Live-Action Discovery Scene	unspecified	95	-	prompt · source
6		Jeweled Scorpion Transforms into a High Heel	unspecified	95	-	prompt · source
7		Seedance Native UGC Vertical Phone Follow-Cam Short	unspecified	95	⚠️ degraded (score 71.2)	prompt · source
8		Seedance 2.5 Photoreal Everyday Short of an Indonesian Girl	unspecified	95	✅ reproduced (score 77.5)	prompt · source
9		Blonde Student Unleashes Superpowers in a High School Hallway	unspecified	94	✅ reproduced (score 81.6)	prompt · source
10		Beach Day Memories Shot on a Smartphone	2.0	93	-	prompt · source
11		Woman Pulls Laughing Man Out Through a Shattered Subway Window	unspecified	93	-	prompt · source
12		Seedance 2.5 Dance Clip Real Enough to Fool You	unspecified	92	⚠️ degraded (score 50.9)	prompt · source
13		POV: Korean Baddie Meets Her Boyfriend in the US	unspecified	92	-	prompt · source
14		Fox’s Self-Filmed Walk Along a Forest Stream	unspecified	92	⚠️ degraded (score 74.5)	prompt · source
15		Hand-Drawn 2D Japanese Bakery Animation	2.0	91	⚠️ degraded (score 55.9)	prompt · source
16		Cinematic Paris Fashion Campaign, Five Shots	2.0	90	✅ reproduced (score 81.8)	prompt · source
17		Seedance 2.0 Cinematic East Asian Lifestyle Short	unspecified	90	✅ reproduced (score 86.8)	prompt · source
18		Boa Hancock Water Obstacle Race Prompt	2.5	89	⚠️ degraded (score 66.3)	prompt · source
19		Meilin-Element Kung Fu Performance	2.0	89	-	prompt · source
20		A Cat's Cozy Day Filmed as a Selfie Vlog	unspecified	89	⚠️ degraded (score 70.2)	prompt · source
21		Seedance Two-Character 2D Anime: Little Kite	unspecified	89	✅ reproduced (score 90.4)	prompt · source
22		A Wizard's Ten-Second Escape Plan: Become a Duck	unspecified	89	-	prompt · source
23		Seedance 2.5 Macro Time-Lapse: Ten Seconds of Germination	unspecified	89	-	prompt · source
24		Erling Haaland Claymation Gardening	2.0	88	✅ reproduced (score 83.3)	prompt · source
25		A Korean Woman’s Summer Day on a Traditional Farm	unspecified	88	-	prompt · source
26		Thirty-Second Tropical Travel Vlog Montage in Bali	unspecified	88	-	prompt · source
27		Realism That Makes Ordinary Life Feel Special	unspecified	88	-	prompt · source
28		A Tiny Bird Emerging from Mud in a Glass Jar	unspecified	88	-	prompt · source
29		Seedance Film It or Run: A First-Person Disaster Short	unspecified	88	✅ reproduced (score 79.3)	prompt · source
30		Handheld Travel Vlog From Apartment to Beach	2.5	87	-	prompt · source

🎬 All Prompts

All 463 cases, with full prompts, live in the gallery under docs/ (sharded so GitHub renders every page). Start from the gallery index, or jump to a version:

• Seedance 2.5 - 56 cases: Full gallery.
• Seedance 2.0 - 130 cases: Part 1 (cases 1–95) · Part 2 (cases 96–130).
• Seedance (version unspecified) - 277 cases: Part 1 (cases 1–91) · Part 2 (cases 92–193) · Part 3 (cases 194–265) · Part 4 (cases 266–277).

🌐 Browse on goodcase.ai

This README is an index. The full experience lives at goodcase.ai: search across every case, the heat leaderboard, the stability ranking, per-case retest logs with output videos, and the installable Skills that grow out of the cases. Every entry here links back to its goodcase.ai record.

Statistics

Metric	Value
Seedance cases in this repo	463
Seedance 2.5	56
Seedance 2.0	130
Seedance (version unspecified)	277
Unique authors	166
Re-run on other models	254 cases / 264 runs
Stability score (measured)	252 cases / avg 77.9
Last updated	2026-09-19
goodcase.ai, all categories	1229 cases / 350 creators
goodcase.ai, AI video	620 cases

Retest batch note: Retest runs carry two MiniMax labels on purpose. The 2026-08 batches (runIds video-minimax-h3-768p-20260809-phase1 / -20260811-top-heat) ran on Flova's MiniMax H3 768p; the 2026-09 batches (video-minimax-h3-768p-20260906-phase1 / -20260907-phase1) ran on fal.ai's minimax/h3-max/text-to-video endpoint, which only existed from 2026-09-03. They are different model tiers on different platforms, so the labels are kept distinct rather than merged. The 2026-08 entries show finalScore: null because a later human-review pass overwrote their structured evidence notes with a prose verdict; those rows carry a human-reviewed verdict instead, and the original judge scores remain in the private eval evidence manifest.

Each case is counted once; a case tagged with several Seedance versions counts under the highest one.

🚀 How to use this repository

1. Start from ⭐ Featured or 🔥 Top 30 and decide what kind of clip you want: vlog, ad, dialogue, action, stylized.
2. Open that category in the 🗂️ Category Overview or the full gallery, read two or three neighbouring cases, and copy the structure first (timeline, shot list, identity lock), then the style words.
3. Install the Skill (npx seedance-prompt-library install) or open the template tables and fill your own subject, setting and beats into the matching template. Check the case's retest verdict before you commit budget to it.

How to Contribute

New prompt cases go through goodcase.ai's review pipeline so provenance and heat score stay verifiable: submit at goodcase.ai/submit (collection standards: goodcase.ai/standards). Prefer GitHub? Open a pull request that adds one JSON file under submissions/ following submissions/TEMPLATE.json; a maintainer pushes it through the same review, and it lands in data/ on the next export.

Pull requests are welcome for template fixes in data/style-library.json, generator and Skill code under scripts/ and agents/, and corrections to English titles or summaries. README.md, README_zh.md, docs/ and the Skill reference are generated from data/, so please don't hand-edit them: change the source, run npm test && npm run generate, and commit the regenerated files in the same PR. See contributing.md for the submission standard, what gets rejected, and how the generator works. This project follows the code of conduct.

🙏 Acknowledgements

This project's format and Skill-packaging approach were shaped by:

• freestylefly/awesome-gpt-image-2 - Template-library + installable-Skill + marketplace pattern.
• YouMind-OpenLab - README-as-gallery with per-entry attribution.
• goodcase.ai - The source of every case, heat score and retest in this repository.

Copyright & Takedown Notice

This repository carries three kinds of material under three different terms.

Code (generator scripts, Agent Skill, tooling) is released under the MIT License, see the LICENSE file. The MIT license covers the code only.

Curation (selection, organization, statistics, template extraction, summaries written by us) is licensed under CC BY 4.0. Reuse it with attribution to awesome-seedance / goodcase.ai.

Prompts and media remain the copyright of their original creators. Prompt text, creator summaries, poster images and video references are quoted from publicly published posts for documentation and study; every entry links back to its original source and to its goodcase.ai record. Nothing here grants a license to the underlying prompt or media beyond what the original post allows.

Takedown process. If you are a rights holder and want an entry removed or corrected, open a GitHub issue with the entry's slug (from its goodcase.ai URL) and the original source link, or contact goodcase.ai directly. Requests are verified against the original post and honored on verification; the entry is removed from data/ and disappears from every generated file on the next regeneration.

Star History

License & Reuse

Code in this repository is open source under the MIT License: use it, modify it, build on it, keep the license notice. Curation is CC BY 4.0; prompts and media stay with their creators. Details above under Copyright & Takedown Notice.