README_pt.md

<div align="center">

<a href="https://evolink.ai/launch/seedance-2-5?utm_source=github&utm_medium=banner&utm_campaign=GPT-Image-2-Seedance-2.5-Workflow"><img src="images/logo.png" alt="GPT Image 2 × Seedance 2.5 Workflow Guide"></a>

[![License: CC BY 4.0](https://img.shields.io/badge/License-CC_BY_4.0-lightgrey.svg)](LICENSE)
[![awesome-seedance-2.5-prompts](https://img.shields.io/badge/📦_awesome--seedance--2.5--prompts-181717?logo=github)](https://github.com/EvoLinkAI/awesome-seedance-2.5-prompts)
[![Seedance-2.5-Gateway-Service](https://img.shields.io/badge/📦_Seedance--2.5--Gateway--Service-181717?logo=github)](https://github.com/EvoLinkAI/Seedance-2.5-Gateway-Service)
[![awesome-seedance-2.5-guide](https://img.shields.io/badge/📦_awesome--seedance--2.5--guide-181717?logo=github)](https://github.com/EvoLinkAI/awesome-seedance-2.5-guide)
[![awesome-gpt-image-2-prompts](https://img.shields.io/badge/📦_awesome--gpt--image--2--prompts-181717?logo=github)](https://github.com/EvoLinkAI/awesome-gpt-image-2-prompts)


[![🇺🇸 English](https://img.shields.io/badge/🇺🇸_English-Default_Source-111111)](README.md)
[![🇪🇸 Español](https://img.shields.io/badge/🇪🇸_Español-Ver-ffb703)](README_es.md)
[![🇵🇹 Português](https://img.shields.io/badge/🇵🇹_Português-Ver-2a9d8f)](README_pt.md)
[![🇯🇵 日本語](https://img.shields.io/badge/🇯🇵_日本語-表示-52b788)](README_ja.md)
[![🇰🇷 한국어](https://img.shields.io/badge/🇰🇷_한국어-보기-4ea8de)](README_ko.md)
[![🇩🇪 Deutsch](https://img.shields.io/badge/🇩🇪_Deutsch-Ansehen-f4a261)](README_de.md)
[![🇫🇷 Français](https://img.shields.io/badge/🇫🇷_Français-Voir-e76f51)](README_fr.md)
[![🇹🇷 Türkçe](https://img.shields.io/badge/🇹🇷_Türkçe-Görüntüle-d62828)](README_tr.md)
[![🇹🇼 繁體中文](https://img.shields.io/badge/🇹🇼_繁體中文-查看-8338ec)](README_zh-TW.md)
[![🇨🇳 简体中文](https://img.shields.io/badge/🇨🇳_简体中文-查看-ef476f)](README_zh-CN.md)
[![🇷🇺 Русский](https://img.shields.io/badge/🇷🇺_Русский-Смотреть-577590)](README_ru.md)

</div>






## 🎬 Introdução

Bem-vindo ao repositório de workflow GPT Image 2 × Seedance 2.5! 🤗

**Reunimos workflows comprovados, templates de prompt e exemplos reais de criadores que combinam GPT Image 2 e Seedance 2.5 para produzir vídeos de IA de alta qualidade.**

GPT Image 2 cuida do "o que desenhar" e da consistência visual. Seedance 2.0 cuida do "como mover" — animando essas imagens em vídeo. Juntos, eles formam um dos pipelines de vídeo com IA mais capazes disponíveis hoje.

A maioria dos casos deste repositório foi curada a partir de criadores no X/Twitter, experimentos da comunidade e workflows reais de produção.

Experimente: [GPT Image 2 + Seedance 2.5](https://evolink.ai/launch/seedance-2-5?utm_source=github&utm_medium=readme&utm_campaign=GPT-Image-2-Seedance-2.5-Workflow)

### Uso rapido

- [GPT Image 2 Gen Skill](https://github.com/EvoLinkAI/gpt-image-2-gen-skill) adiciona geracao e edicao de imagens com GPT Image 2 ao seu agente
- [Seedance 2 Video Gen Skill](https://github.com/EvoLinkAI/seedance2-video-gen-skill-for-openclaw) adiciona geracao de video com Seedance 2 ao seu agente

### Instalacao rapida

#### Executar direto no Terminal

```bash
npx evolink-gpt-image@latest
npx evolink-seedance@latest
```

Se voce mesmo for instalar manualmente no Terminal, pode executar essas duas linhas diretamente. Os instaladores tentam detectar automaticamente um diretorio de skills conhecido e, se nao encontrarem, perguntam onde instalar.

#### Claude Code

```bash
npx evolink-gpt-image@latest -y --path ~/.claude/skills
npx evolink-seedance@latest -y --path ~/.claude/skills
```

#### OpenCode

```bash
npx evolink-gpt-image@latest -y --path ~/.opencode/skills
npx evolink-seedance@latest -y --path ~/.opencode/skills
```

#### Cursor

```bash
npx evolink-gpt-image@latest -y --path .cursor/skills
npx evolink-seedance@latest -y --path .cursor/skills
```

#### OpenClaw

```bash
npx evolink-gpt-image@latest -y
npx evolink-seedance@latest -y
```

A versao de Terminal acima serve para instalacao manual. Os comandos especificos por host abaixo sao melhores quando voce quer que um AI Agent instale os skills sem fazer perguntas adicionais.

Se isto for útil para você, considere dar uma estrela. ⭐



## 📰 Notícias

- **26 de maio de 2026:** Adicionados 30 novos trabalhos à [🌟 Vitrine da Comunidade](#-vitrine-da-comunidade) do lote de 17 a 23 de maio (slugs 104–133)
- **2 de maio de 2026:** Adicionada [🌟 Vitrine da Comunidade](#-vitrine-da-comunidade) — 70 trabalhos mais recentes em GPT Image 2 × Seedance 2.5 de criadores no X (29 de abril – 2 de maio)

## 📑 Menu

- [🎬 Introdução](#-introdução)
- [📰 Notícias](#-notícias)
- [📑 Menu](#-menu)
- [🎥 Técnicas de Storyboard](#-técnicas-de-storyboard)
  - [Caso 1: Storyboard padrão → Vídeo (por @kiyoshi_shin)](#caso-1-storyboard-padrão--vídeo-por-kiyoshi_shin)
  - [Caso 2: Método de storyboard em grade 3×3 (por @servasyy_ai)](#caso-2-método-de-storyboard-em-grade-33-por-servasyy_ai)
  - [Caso 10: Referência multiframe → Vídeo de corte rápido (por @heygentlewhale)](#caso-10-referência-multiframe--vídeo-de-corte-rápido-por-heygentlewhale)
  - [Caso 19: Controle de custo com storyboard primeiro (por @0xbisc)](#caso-19-controle-de-custo-com-storyboard-primeiro-por-0xbisc)
- [📱 Comercial & Produto](#-comercial--produto)
  - [Caso 5: Vídeo demo de MVP de app (por @Shin_Engineer)](#caso-5-vídeo-demo-de-mvp-de-app-por-shin_engineer)
  - [Caso 6: Comercial de 15 segundos (por @ai_mitosan)](#caso-6-comercial-de-15-segundos-por-ai_mitosan)
  - [Caso 15: Comercial de luxo — Do storyboard ao filme (por @insmind_com)](#caso-15-comercial-de-luxo--do-storyboard-ao-filme-por-insmind_com)
  - [Caso 16: Vídeo gastronômico cinematográfico (por @kingofdairyque)](#caso-16-vídeo-gastronômico-cinematográfico-por-kingofdairyque)
  - [Caso 26: Imagem de produto → Anúncio em vídeo curto (por @insmind_com)](#caso-26-imagem-de-produto--anúncio-em-vídeo-curto-por-insmind_com)
- [🎨 Animação & Personagem](#-animação--personagem)
  - [Caso 3: Folha de personagem → Animação (por @YaReYaRu30Life)](#caso-3-folha-de-personagem--animação-por-yareyaru30life)
  - [Caso 4: Vídeo em estilo de abertura de anime (por @Toshi_nyaruo_AI)](#caso-4-vídeo-em-estilo-de-abertura-de-anime-por-toshi_nyaruo_ai)
  - [Caso 12: Claude Code × Folha de personagem → Animação (por @old_pgmrs_will)](#caso-12-claude-code--folha-de-personagem--animação-por-old_pgmrs_will)
  - [Caso 13: Grade de sequência de dança → Vídeo de dança (por @Ciri_ai)](#caso-13-grade-de-sequência-de-dança--vídeo-de-dança-por-ciri_ai)
  - [Caso 14: Página de quadrinhos → Vídeo animado (por @nimentrix)](#caso-14-página-de-quadrinhos--vídeo-animado-por-nimentrix)
  - [Caso 25: Coreografia K-Pop — Controle detalhado (por @Kashberg_0)](#caso-25-coreografia-k-pop--controle-detalhado-por-kashberg_0)
  - [Caso 27: Animação de introdução de personagem (por @0xbisc)](#caso-27-animação-de-introdução-de-personagem-por-0xbisc)
- [🎵 Videoclipe & Curta-metragem](#-videoclipe--curta-metragem)
  - [Caso 7: Videoclipe com Suno (por @fukaborichannel)](#caso-7-videoclipe-com-suno-por-fukaborichannel)
  - [Caso 8: Curta em estilo cyberpunk (por @ponyodong)](#caso-8-curta-em-estilo-cyberpunk-por-ponyodong)
  - [Caso 11: MV japonês — Toolchain completo de IA (por @Tz_2022)](#caso-11-mv-japonês--toolchain-completo-de-ia-por-tz_2022)
  - [Caso 20: Shotlist Claude → Videoclipe (por @CoffeeVectors)](#caso-20-shotlist-claude--videoclipe-por-coffeevectors)
- [🎮 Conceito de Jogo](#-conceito-de-jogo)
  - [Caso 9: Jogos e conteúdo interativo (por @op7418)](#caso-9-jogos-e-conteúdo-interativo-por-op7418)
  - [Caso 17: Animação de interface de jogo — Pipeline completo (por @0xInk_)](#caso-17-animação-de-interface-de-jogo--pipeline-completo-por-0xink_)
  - [Caso 24: Conceito de jogo de cidade estilo GTA (por @markgadala)](#caso-24-conceito-de-jogo-de-cidade-estilo-gta-por-markgadala)
- [🛠 Ferramentas de Produção](#-ferramentas-de-produção)
  - [Caso 18: Workflow automatizado com agente único (por @venturetwins)](#caso-18-workflow-automatizado-com-agente-único-por-venturetwins)
  - [Caso 21: Grade de casting — Audição de atores (por @8fstudioz)](#caso-21-grade-de-casting--audição-de-atores-por-8fstudioz)
  - [Caso 22: Escultura 3D → Render IA → Animação (por @_DAntunes_)](#caso-22-escultura-3d--render-ia--animação-por-_dantunes_)
- [🌟 Vitrine da Comunidade](#-vitrine-da-comunidade)
- [💡 Dicas & Técnicas](#-dicas--técnicas)
  - [Guia de consistência](#guia-de-consistência)
  - [Templates de prompt](#templates-de-prompt)
- [🚀 Experimente no Evolink](#-experimente-no-evolink)
- [🙏 Agradecimentos](#-agradecimentos)


## 🎥 Técnicas de Storyboard

<!-- Case 1: Standard Storyboard → Video (by @kiyoshi_shin) -->
### Caso 1: [Storyboard padrão → Vídeo](https://x.com/kiyoshi_shin/status/2047133524403400847) (por [@kiyoshi_shin](https://x.com/kiyoshi_shin))

O workflow mais comum. Use o GPT Image 2 para gerar um painel de storyboard e depois anime-o com o Seedance 2.0. Ideal para vídeos promocionais, dramas curtos e aberturas de animação.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/storyboard_case1/input.jpg" width="280" alt="Storyboard gerado pelo GPT Image 2"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/ac25fc3d-b6cb-4149-a8ba-e7e10c5b1faa" width="280" controls></video></td>
</tr></table>

**Passos:**

1. Descreva sua cena para o GPT Image 2 e gere uma imagem de storyboard
2. Importe o storyboard para o Seedance 2.0 usando o modo Image-to-Video
3. Exporte cada clipe e monte tudo no seu software de edição

**Prompt para GPT Image 2:**

```
Create a 6-panel storyboard for a 15-second brand promotional video. Label each panel with a shot description.
Style: cinematic, cool color tone, widescreen 16:9.
Content: the journey of a product from factory to the customer's hands.
```

**Prompt para Seedance 2.0:**

```
Cinematic brand advertisement, slow camera push-in, product centered in frame, warm side lighting, soft background blur, no people, 3 seconds.
```

> [!NOTE]
> Gere imagens de storyboard em 16:9 para evitar o recorte automático do Seedance. Defina a taxa de quadros em 24 fps para seguir o padrão cinematográfico. Mantenha cada painel simples — quanto mais simples o conteúdo, mais preciso tende a ser o movimento gerado.


<!-- Case 2: 3×3 Grid Storyboard Method (by @servasyy_ai) -->
### Caso 2: [Método de storyboard em grade 3×3](https://x.com/servasyy_ai/status/2047198012750143999) (por [@servasyy_ai](https://x.com/servasyy_ai))

Uma técnica importante descoberta pela comunidade: compor todos os painéis do storyboard em uma única imagem em grade 3×3 antes de importar para o Seedance reduz bastante a taxa de falha em comparação com importar quadro por quadro.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/storyboard_case2/output.jpg" width="400" alt="Storyboard em grade 3×3 gerado"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/00f32388-a17b-4b9c-8da3-1956436ce91b" width="400" controls></video></td>
</tr></table>



**Passos:**

1️⃣ Insira o conteúdo que deseja criar + o ✅ prompt "Create a storyboard in a 3×3 grid format"
2️⃣ Crie um prompt a partir da imagem do passo 1 com o ChatGPT
3️⃣ Referencie a imagem do passo 1 no Seedance
4️⃣ Insira o prompt criado no passo 2.

**Prompt para GPT Image 2:**

```
[describe your scene] and Create a storyboard in a 3×3 grid format
```

**Prompt para Seedance 2.0:**
transforme esta imagem em vídeo
```
[Describe the motion and style. Example: Japanese full-color animation, fast cuts, high frame count, 24fps, dark fantasy anime OP style, intense battle scenes.]
```

> [!NOTE]
> **Substitua o conteúdo entre colchetes antes de usar.** Este método funciona porque o Seedance analisa a intenção de movimento a partir de uma única imagem. A grade fornece referência direcional e produz um movimento mais coerente do que imagens separadas.


<!-- Case 10: Multi-Frame Reference Storyboard (by @heygentlewhale + @ai_gezgini) -->
### Caso 10: [Referência multiframe → Vídeo de corte rápido](https://x.com/heygentlewhale/status/2047969137969004946) (por [@heygentlewhale](https://x.com/heygentlewhale))

Forneça ao Seedance 2.0 uma imagem de storyboard com vários quadros de referência e instrua-o a seguir a ordem da sequência. O modelo lê as posições dos quadros como sinais de cena e gera uma edição de corte rápido coerente — sem necessidade de montar clipes manualmente.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/storyboard_case10/input.jpg" width="280" alt="Storyboard multiframe do GPT Image 2"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/4d7af334-4e49-4de4-899e-803f72116c21" width="280" controls></video></td>
<td align="center"><video src="https://github.com/user-attachments/assets/5def7e00-6fc6-4a4e-8075-5f37cb24b84c" width="280" controls></video></td>
</tr></table>

**Passos:**

1. Gere no GPT Image 2 uma imagem de storyboard multipanel (12 quadros, grade 3×4 ou 4×3)
2. Carregue o storyboard como imagem de referência no Seedance 2.0
3. Escreva um prompt de sequenciamento que indique o número de quadros e o estilo de edição

**Prompt para GPT Image 2:**

```
Create a 12-panel storyboard grid for a [N]-second [genre] film:
- 4 columns × 3 rows, left-to-right, top-to-bottom reading order
- Each panel: [shot type] + [action description]
- Location: [setting], Time: [day/night], Mood: [atmosphere]
- Consistent character design and scene across all panels
- No text labels, no panel borders
Output as a single image.
```

**Prompt para Seedance 2.0:**

```
Follow the storyboard sequence of the 12 reference frames in image1, edited as a fast-cut memory montage.
[Describe visual style — example below:]
A nostalgic romance film set in 1990s Singapore, shot on 35mm film in Kodak Portra 800 style.
Soft grain, dreamy blur, warm highlights, and slight color shifts create a vintage cinematic atmosphere.
```

**Prompt de sequenciamento universal (via [@ai_gezgini](https://x.com/ai_gezgini/status/2047349122315805016)):**

```
Use this storyboard to generate a video, follow the scene order, keep transitions smooth,
and preserve cinematic lighting and pacing.
[Add any extra visual details you want.]
```

> [!NOTE]
> Este prompt funciona em qualquer gênero — substitua a descrição de estilo por ficção científica, terror, documentário ou qualquer outro. A frase-chave é `follow the storyboard sequence of the [N] reference frames` — ela instrui o Seedance a tratar as posições dos quadros como uma linha do tempo em vez de uma composição única.


<!-- Case 19: Storyboard-First Cost Control (by @0xbisc) -->
### Caso 19: [Controle de custo com storyboard primeiro](https://x.com/0xbisc/status/2049100073481716076) (por [@0xbisc](https://x.com/0xbisc))

Uma abordagem voltada para produção: itere no storyboard com o GPT Image 2 primeiro (barato), e só gere o vídeo quando a composição estiver travada (caro). Isso economiza créditos significativos, pois iterações de vídeo consomem de 10 a 50 vezes mais do que iterações de imagem. 298 curtidas / 13K visualizações / 291 salvamentos.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/storyboard_case19/output.jpg" width="400" alt="Saída do workflow storyboard primeiro"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/09e04d80-c0d1-4a8c-9b74-2efe474acfcd" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Gere uma grade de storyboard com 8 painéis no GPT Image 2
2. Revise cada painel — regenere ou edite painéis individuais até ficar satisfeito
3. Somente quando o storyboard completo estiver travado, importe para o Seedance 2.0
4. Gere o vídeo de uma só vez a partir do storyboard finalizado

**Prompt para GPT Image 2:**

```
Create a single cinematic storyboard image containing 8 panels, arranged in a 4-column horizontal grid layout across the canvas.
Panels are evenly distributed in 4 columns, forming a balanced multi-row composition.
Use generous white space between panels and around the entire layout, creating a calm, refined editorial presentation.
Minimalist editorial design, white background, precise alignment, consistent spacing, strong grid system.
Subtle divider lines, ultra-thin, low-contrast, neutral tone, strictly aligned to the grid.
Each panel is presented as a clean modular card with a clear hierarchy: image frame on top, minimal text block below.
Refined modern sans-serif typography, light to regular weight, tight typographic control, consistent scale rhythm.
Visual consistency across all panels: a white flying dragon and a short blond-haired young male wearing a loose flowing white robe riding on its back. A glowing spherical object remains consistent in appearance.
Style: live-action cinematic realism, human actor proportions, natural skin detail, physically accurate lighting, real-world materials, high-end film still quality, ultra high resolution, sharp focus.
Cinematic sequence:
Scene 01
Shot type
Wide shot, low-angle tracking
Narrative
Dragon skims rapidly above ground toward vast geometric ruins, rider standing steadily, bright daylight
Scene 02
Shot type
Wide shot, side view
Narrative
Dragon enters ruin airspace, stone structures begin sinking and shifting, geometry reconfigures
Scene 03
Shot type
Medium shot, tracking
Narrative
Dragon navigates through moving structures forming a spatial maze, rider leans forward, focused
Scene 04
Shot type
Push-in shot, centered composition
Narrative
Massive floating ring appears at center, glowing sphere suspended inside, focal point established
Scene 05
Shot type
Wide-angle, low-angle tracking, high speed
Narrative
Dragon dives through narrow moving gaps, massive structures pass extremely close, intense speed and compression
Scene 06
Shot type
Close-up, centered composition
Narrative
Rider reaches forward to grasp glowing sphere, golden light illuminating face and arm
Scene 07
Shot type
Wide shot, upward motion
Narrative
Dragon ascends sharply as ruins collapse below, structures sinking and closing, dust fills air
Scene 08
Shot type
Wide shot, high-angle view
Narrative
Dragon exits above collapsed ruins, open sky, clear silhouette of rider and dragon, resolution moment
```

**Prompt para Seedance 2.0:**

```
Generate video based strictly on storyboard @ image1. Follow the storyboard exactly as shown, matching each panel's composition, framing, and action. Keep perfect visual continuity with no errors or inconsistencies.
```

> [!NOTE]
> **Por que storyboard primeiro economiza custos:** Iterações de vídeo consomem créditos rapidamente. Com um storyboard, você pode ajustar plano a plano e detectar problemas cedo. A etapa de vídeo se torna uma única renderização final em vez de um loop caro de tentativa e erro. O feedback da comunidade confirma que este é o workflow mais eficiente em termos de orçamento para sequências longas.


## 📱 Comercial & Produto

<!-- Case 5: App MVP Demo Video (by @Shin_Engineer) -->
### Caso 5: [Vídeo demo de MVP de app](https://x.com/Shin_Engineer/status/2047182050323812381) (por [@Shin_Engineer](https://x.com/Shin_Engineer))

Use o GPT Image 2 para gerar screenshots de interface com aparência final de um app que ainda não existe, e depois anime-os com o Seedance 2.0 para criar um demo de produto. Publique no TikTok ou em redes sociais para testar o interesse do mercado antes de desenvolver.

| Saída |
| :----: |
| <a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/app_case5/output0.jpg" width="400" alt="Screenshot 1 de UI de app gerado pelo GPT Image 2"></a> |

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/app_case5/output1.jpg" width="220" alt="App UI screenshot 2"></a></td>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/app_case5/output2.jpg" width="220" alt="App UI screenshot 3"></a></td>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/app_case5/output3.jpg" width="220" alt="App UI screenshot 4"></a></td>
</tr></table>

**Passos:**

1. Descreva seu conceito de app e sua linguagem de design para o GPT Image 2
2. Gere de 3 a 5 screenshots principais da interface (home, funcionalidade, perfil)
3. Ordene as imagens no fluxo do usuário e importe para o Seedance 2.0
4. Exporte o vídeo demo e publique para testar a reação do mercado

**Prompt para GPT Image 2:**

```
Design [N] UI screenshots for a "[app concept]" app:
1. [Page 1 name and description]
2. [Page 2 name and description]
3. [Page 3 name and description]
Style: [iOS/Android] native design language, [primary color] accent, [light/dark] mode.
Output as realistic app screenshots, not wireframes or mockups.
```

**Prompt para Seedance 2.0:**

```
Smooth app UI transition animation, screen tap interaction, natural interface motion, clean and modern feel, 3 seconds.
```

> [!NOTE]
> **Substitua os placeholders entre colchetes antes de usar.** Na legenda do vídeo, não o identifique como gerado por IA — publique como demo de produto e observe o feedback real do público nos comentários.


<!-- Case 6: 15-Second Commercial (by @ai_mitosan) -->
### Caso 6: [Comercial de 15 segundos](https://x.com/ai_mitosan/status/2047146600422846762) (por [@ai_mitosan](https://x.com/ai_mitosan))

Workflow em duas etapas: o GPT Image 2 gera a imagem principal e o storyboard correspondente; depois o Seedance 2.0 anima cada clipe. Monte tudo com legendas e música para obter uma peça completa de 15 segundos.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/commercial_case6/input1.jpg" width="280" alt="Imagem principal de produto do GPT Image 2"></a></td>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/commercial_case6/input2.jpg" width="280" alt="Storyboard do GPT Image 2"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/09ae3c57-b8fb-4323-ba76-7777541fe4a3" width="280" controls></video></td>
</tr></table>

**Passos:**

1. Gere a imagem principal + storyboard com o GPT Image 2
2. Alimente no Seedance 2.0 para transformar em vídeo

**Guia de quantidade de painéis:**

| Duração do vídeo | Painéis | Duração por clipe |
| :---: | :---: | :---: |
| 15 segundos | 4–5 | 3–4 segundos |
| 30 segundos | 8–10 | 3 segundos |
| 60 segundos | 15–18 | 3–4 segundos |

**Prompt para GPT Image 2:**

```
夜カフェ　深夜スイーツをテーマにした15秒CMを作るので、絵コンテを作って。 
プロの映像クリエイターによる15秒、８カット、マルチショット、ライティング重視。
```

**Prompt para Seedance 2.0:**

```
15秒、８カット、マルチショット、ライティング重視
```


<!-- Case 15: Luxury Commercial — Storyboard to Film (by @insmind_com) -->
### Caso 15: [Comercial de luxo — Do storyboard ao filme](https://x.com/insmind_com/status/2049481439285223785) (por [@insmind_com](https://x.com/insmind_com))

Gere uma grade de storyboard 3×4 (12 quadros) para um anúncio de fragrância de luxo com o GPT Image 2, depois anime em um filme cinematográfico de nível de marca com o Seedance 2.0. O fluxo estruturado — introdução → ritual → transformação → resolução → fechamento de marca — produz um arco narrativo completo em uma única geração. 371 curtidas / 84K visualizações / 255 salvamentos.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/luxury_case15/output.jpg" width="400" alt="Storyboard de comercial de luxo"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/281fef1e-f42d-442c-b06e-44d7cff221ec" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Gere uma grade de storyboard com 12 quadros (3×4) no GPT Image 2, usando layout editorial e estética de marca de luxo
2. Importe a grade de storyboard como uma única imagem de referência no Seedance 2.0
3. Instrua o Seedance a animar a sequência como um comercial cinematográfico de luxo
4. Adicione música e faça a correção de cor final no seu software de edição

**Prompt para GPT Image 2:**

```
Create a high-end 9:16 luxury fragrance pitch deck storyboard in 3x4 grid (12 frames), editorial layout, Aesop/Byredo style, beige + lavender palette. Structured flow: intro → ritual → transformation → resolution → brand closure. Each frame split: top cinematic image (no text) + bottom storyboard notes. Luxury minimal European aesthetic, calm femininity, slow living mood. Candle is the emotional center throughout
```

**Prompt para Seedance 2.0:**

```
Animate the provided 3x4 storyboard into a smooth cinematic video. Preserve exact shot order and continuity. Use slow push-in, gentle pan, subtle orbit, and rack focus. Lighting transitions from soft morning glow to warm golden ambient light. Fragrance brand editorial aesthetic, minimal luxury, soft film grain. No new shots, no reordering, candle remains emotional focus in all scenes
```

> [!NOTE]
> O layout editorial de pitch-deck (com notas de direção visual em cada quadro) fornece ao Seedance pistas narrativas mais fortes do que uma grade simples. Este workflow se adapta a qualquer categoria de produto de luxo — skincare, relógios, moda — basta trocar a paleta e as referências de produto.

<!-- Case 16: Cinematic Food Video (by @kingofdairyque) -->
### Caso 16: [Video gastronômico cinematográfico](https://x.com/kingofdairyque/status/2049812014596599834) (por [@kingofdairyque](https://x.com/kingofdairyque))

Use GPT Image 2 + Seedance 2.5 para criar vídeos ultra-realistas de preparação de alimentos com descrições de planos cronometrados. Cada segmento de tempo (0–2s, 2–4s, etc.) define um ângulo de câmera e uma ação específicos, dando ao Seedance controle preciso sobre a sequência de 15 segundos. 55 curtidas / 1K visualizações.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/food_case16/input.jpg" width="400" alt="Entrada de storyboard de vídeo gastronômico"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/30a20e57-8384-4117-adf7-4f92faebeb33" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Escreva uma lista de planos cronometrada detalhada descrevendo cada segmento de 2 segundos
2. Gere a imagem de storyboard com o GPT Image 2 baseada na lista de planos
3. Alimente a imagem + o prompt cronometrado completo no Seedance 2.0
4. O modelo segue a estrutura de timestamps para produzir uma sequência de culinária coerente

**Prompt para Seedance 2.0:**

```
Ultra-realistic cinematic 15-second vertical video (9:16), 4K, dark rustic wooden table,
soft dramatic lighting, shallow depth of field, natural textures, no text, no subtitles, no music.
[0-2s] Top-down shot: hands gently touch a bowl of [ingredient]. Subtle ambient movement.
[2-4s] Side close shot: [preparation action] in slow motion, catching warm light.
[4-6s] Macro shot: hands [action]. Texture detail visible.
[6-8s] 45-degree angle: [ingredient] pours into [vessel]. Natural bounce and movement.
[8-10s] Top angled close-up: hands carefully [assembly action]. Precise controlled motion.
[10-12s] Side shot: oven door opens. Warm golden light spills out with gentle steam.
[12-14s] Close-up: [finishing touch]. Surface becomes glossy, light reflecting softly.
[14-15s] Final frontal shot: finished [dish] on rustic table. Hands enter frame softly.
```

> [!NOTE]
> A técnica de prompt cronometrado dá ao Seedance uma linha do tempo precisa plano a plano. Funciona para qualquer vídeo de produto ou processo — unboxing, artesanato, preparo de coquetéis. Mantenha cada segmento em 2 segundos e descreva tanto o ângulo de câmera quanto a ação para melhores resultados.

<!-- Case 26: Product Image → Short Video Ad (by @insmind_com) -->
### Caso 26: [Imagem de produto → Anúncio em vídeo curto](https://x.com/insmind_com/status/2049843814337306974) (por [@insmind_com](https://x.com/insmind_com))

Transforme imagens estáticas de produtos em vídeos que chamam atenção nas redes sociais. Carregue suas fotos de produto existentes como referência no GPT Image 2, gere uma composição de cena e depois anime com o Seedance 2.0. Projetado para e-commerce e marketing em redes sociais — produza conteúdo pronto para TikTok/Reels a partir de fotos de produto que você já possui.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/product_case26/output.jpg" width="400" alt="Saída de anúncio em vídeo de produto"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/880c0019-e45a-4eb9-be6f-638ff71a0e0f" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Insira as imagens do produto.
2. Gere a cena principal.
3. Defina o movimento e a estrutura.
4. Estilo e restrições

> [!NOTE]
> Isso difere do Caso 15 (comercial de luxo) pois começa a partir de fotos de produto existentes em vez de gerar tudo do zero. Ideal para vendedores de e-commerce que já possuem imagens de produto e querem convertê-las em anúncios em vídeo rapidamente.

## 🎨 Animação & Personagem

<!-- Case 3: Character Sheet → Animation (by @YaReYaRu30Life) -->
### Caso 3: [Folha de personagem → Animação](https://x.com/YaReYaRu30Life/status/2047203375314571501) (por [@YaReYaRu30Life](https://x.com/YaReYaRu30Life))

Gere uma folha de personagem com três vistas (frente, lateral, costas) com o GPT Image 2, depois use-a como âncora para animação no Seedance 2.0. Ideal para personagens de anime, personagens de jogos e revelações de figuras.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/character_case3/input0.jpg" width="260" alt="Folha de personagem frente"></a></td>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/character_case3/input1.jpg" width="260" alt="Folha de personagem lateral"></a></td>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/character_case3/input2.jpg" width="260" alt="Folha de equipamentos"></a></td>
</tr></table>

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/character_case3/output.jpg" width="400" alt="Folha de personagem combinada com equipamentos"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/92a0aa56-441f-40db-b9c9-13410254cb3f" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Desenho de três vistas (personagem) + dois desenhos de três vistas de equipamentos. Com base nisso, prepare desenhos de três vistas com cada equipamento equipado em uma única imagem. Por questões de contagem de imagens na postagem, o anexo do personagem foi omitido
2. Crie um storyboard baseado neste desenho de três vistas
3. Transforme o storyboard em vídeo usando o Seedance 2.0

**Prompt para GPT Image 2:**

```
Create a storyboard based on this three-view drawing  
```

**Prompt para Seedance 2.0:**

```
Turn the storyboard into video using Seedance2.0
```

<!-- Case 4: Anime OP Style Video (by @Toshi_nyaruo_AI) -->
### Caso 4: [Vídeo em estilo de abertura de anime](https://x.com/Toshi_nyaruo_AI/status/2047216971184546231) (por [@Toshi_nyaruo_AI](https://x.com/Toshi_nyaruo_AI))

Use o GPT Image 2 para criar uma imagem de ambientação de cena, depois deixe o Seedance 2.0 animar livremente. Comparar saídas controladas (guiadas por storyboard) e livres (apenas prompt) ajuda a decidir a abordagem certa para cada plano.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/character_case4/output0.jpg" width="280" alt="Saída de abertura de anime 1"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/f08a2fee-89a7-4c7c-a58a-f1306f87419a" width="280" controls></video></td>
<td align="center"><video src="https://github.com/user-attachments/assets/09d81a41-b5c5-47f3-8c67-442b7a93b019" width="280" controls></video></td>
</tr></table>

**Passos:**

1. O Grok criou letras para uma abertura de anime fictícia
2. Usou o GPT Image 2 para transformar as letras em um storyboard
3. Usou o Seedance 2.0 para gerar os vídeos

**Prompt para GPT Image 2:**

```
turn the lyrics into a storyboard
```

**Prompt para Seedance 2.0:**

```
Japanese full-color anime, fast cuts, high frame count, 24fps. Dark fantasy anime OP style. Epic battle between protagonist and massive supernatural creatures. High-impact sequence of scenes. Only [character name] appears.
```

> [!NOTE]
> Quando o Seedance anima livremente (sem referência de storyboard), os resultados podem ser mais dinâmicos, mas menos consistentes com a imagem de origem. Use controle por storyboard para planos-chave de personagem e animação livre para sequências de ação.

<!-- Case 12: Claude Code + Character Sheet → Animation (by @old_pgmrs_will) -->
### Caso 12: [Claude Code × Folha de personagem → Animação](https://x.com/old_pgmrs_will/status/2045091769180914019) (por [@old_pgmrs_will](https://x.com/old_pgmrs_will))

Use o Claude Code para escrever a construção de mundo e o lore do personagem, depois passe descrições estruturadas para o GPT Image 2 gerar o visual-chave do personagem, e então anime com o Seedance 2.0. Workflow amigável para desenvolvedores na criação de IPs originais. 191 curtidas / 7K visualizações.

<table><tr>
<td align="center"><a href="https://evolink.ai/seedance2?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/character_case12/output.jpg" width="400" alt="Visual-chave de personagem Claude Code + GPT Image 2"></a></td>
</tr></table>

**Passos:**

1. Use o Claude Code para redigir notas de construção de mundo e uma especificação estruturada de personagem (nome, aparência, personalidade, cenário)
2. Alimente a especificação do personagem diretamente no GPT Image 2 para gerar um visual-chave ou folha de personagem
3. Use o visual-chave como imagem de referência no Seedance 2.0 e anime

> [!NOTE]
> O Claude Code gera texto estruturado — especificações de personagem, descrições de cena, roteiros de diálogo — que o GPT Image 2 processa bem como prompts detalhados. Este pipeline é particularmente eficaz para IPs de histórias originais: construa o lore em código, visualize no GPT Image 2, anime no Seedance.

<!-- Case 13: Dance Sequence Grid → Dance Video (by @Ciri_ai) -->
### Caso 13: [Grade de sequência de dança → Vídeo de dança](https://x.com/Ciri_ai/status/2049034340160704643) (por [@Ciri_ai](https://x.com/Ciri_ai))

Gere uma grade 4×4 de poses de dança com o GPT Image 2, depois alimente no Seedance 2.0 como referência de movimento. O modelo lê a grade como uma sequência coreográfica e gera um vídeo de dança contínuo. Variante avançada: carregue múltiplas referências de personagem para transições de roupa sincronizadas com a batida. 161 curtidas / 9K visualizações.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/dance_case13/output.jpg" width="400" alt="Saída da grade de sequência de dança"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/39376245-e7c7-4812-b770-9e81acf4eca2" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Gere uma imagem em grade 4×4 mostrando um personagem em poses de dança sequenciais com o GPT Image 2
2. Carregue a grade como imagem de referência no Seedance 2.0
3. Instrua o Seedance a seguir a sequência de dança da imagem de referência
4. (Avançado) Carregue o personagem com Roupa A, personagem com Roupa B e a grade de dança como três referências para transições de roupa durante a dança

**Prompt para GPT Image 2:**

```
Transform the input image into a stylized K-pop dance tutorial poster with a fashion-forward streetwear aesthetic, keeping the exact 4x4 grid layout (16 panels) and choreography structure.
Core Composition
Maintain a 16-panel grid (4 columns × 4 rows) with clean spacing

Each panel shows the same female dancer performing sequential choreography

Preserve panel numbering (1–16) in bold, modern UI-style labels

Keep step titles and instructional captions, but redesign typography to feel like K-pop album graphics / dance practice overlays

Subject Styling (K-pop Idol Inspired)

Young female dancer with soft glam K-pop makeup (dewy skin, subtle shimmer, defined eyes)

Hair: long, sleek, slightly dynamic (motion-friendly, flowing during moves)

Expression: confident, charismatic stage presence

Outfit (streetwear-inspired):

Cropped hoodie or oversized zip-up jacket

Cargo pants or parachute pants with straps

Chunky sneakers or platform boots

Optional accessories: chain necklace, ear cuffs, fingerless gloves

Visual Style

Switch from plain grayscale → high-contrast + soft neon accents

Base palette: black, white, gray

Accent colors: neon pink, electric blue, or violet glow (subtle, not overpowering)

Lighting:

Studio lighting with a soft glow + rim light effect

Slight stage-light vibe, like a K-pop dance practice video

Graphics & Effects

Add dynamic motion trails and glow accents on arms, legs, and hair movement

Replace basic arrows with stylized motion graphics (neon strokes, swooshes)

Subtle light streaks or particle effects for energy

Optional faint floor reflection or glossy surface

Typography

Titles: bold, modern, slightly condensed sans-serif (K-pop album style)

Add subtle glow or gradient to titles

Instruction text: clean, minimal, slightly futuristic UI style

Panel numbers: inside rounded squares or pill shapes with neon outline

Camera & Framing

Full-body framing in each panel (consistent scale)

Straight-on angle, but with slight dynamic tilt or perspective energy

Maintain clarity of movement for instructional purpose

Mood & Energy

Feels like a K-pop dance practice meets fashion editorial

Clean but energetic

Stylish, rhythmic, performance-driven

Important Constraints

Keep choreography readable and sequential

Do NOT merge panels or change layout

Maintain consistency of dancer identity across all panels
```

**Prompt para Seedance 2.0:**

```
Character from Image 1 performs the dance based on the breakdown in Image 3. Midway through the performance, they switch outfits on beat into the character from Image 2. Then, the character from Image 2 continues and completes the remaining dance steps from Image 3. Emphasize precise beat synchronization with the music
```

> [!NOTE]
> Esta técnica funciona para qualquer sequência de movimento — dança, artes marciais, esportes. A grade 4×4 fornece ao Seedance 16 quadros de referência para interpolar, produzindo um movimento mais suave do que com menos painéis.
>
> **Variantes da comunidade:** [@airina_xyz](https://x.com/airina_xyz/status/2049830199236190326) demonstrou o workflow básico com um dançarino de rua urbano. [@Kashberg_0](https://x.com/Kashberg_0/status/2049697925262102689) usou pranchas de personagem + quadros de referência de movimento para coreografia K-Pop (52 curtidas / 2K visualizações).

<!-- Case 14: Comic Page → Animated Video (by @nimentrix) -->
### Caso 14: [Página de quadrinhos → Vídeo animado](https://x.com/nimentrix/status/2049560412979708334) (por [@nimentrix](https://x.com/nimentrix))

Crie uma página de quadrinhos multipanel com o GPT Image 2 — layout diagonal, balões de fala, narrativa cinematográfica — depois anime a página inteira em vídeo com o Seedance 2.0. O modelo lê os painéis do quadrinho como uma sequência narrativa e produz um curta animado contínuo. 330 curtidas / 21K visualizações / 360 salvamentos.

<table><tr>
<td align="center"><strong>Entradas GPT Image 2</strong><br><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/comic_case14/input1.jpg" width="260" alt="Referência de personagem 1 — dragão"></a></td>
<td align="center"><br><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/comic_case14/input2.jpg" width="260" alt="Referência de personagem 2 — garoto"></a></td>
<td align="center"><strong>Entrada Seedance 2.0</strong><br><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/comic_case14/input3.jpg" width="260" alt="Página de quadrinhos gerada pelo GPT Image 2"></a></td>
</tr></table>

<table><tr>
<td align="center"><video src="https://github.com/user-attachments/assets/0b5038e2-dfca-4c65-b5d7-a719a74408b0" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Crie uma folha de design de personagem (vistas frontal, lateral, traseira) com o GPT Image 2 para travar a aparência do personagem
2. Gere uma página de quadrinhos multipanel usando o personagem como referência
3. Importe a página de quadrinhos no Seedance 2.0 e anime

**Prompt para GPT Image 2 — Folha de personagem:**

```
Create a character design style sheet for [describe your character]:
front view, side view, back view on white background.
Make the aspect ratio 4:3.
```

**Prompt para GPT Image 2 — Página de quadrinhos:**

```
[Character description] and [companion], american comic multi-panel illustration,
diagonal layout, six panels, cinematic storytelling, clear reading flow, with speech bubbles.
[Describe the story sequence across panels.]
```

**Prompt para Seedance 2.0:**

```
Animate this comic page as a cinematic sequence. Follow the panel order from top-left to bottom-right.
Smooth transitions between panels, maintain character consistency, cinematic camera movement.
```

> [!NOTE]
> O layout diagonal e os balões de fala fornecem ao Seedance pistas visuais claras para os limites dos painéis e a ordem de leitura. Para melhores resultados, mantenha a ação de cada painel simples e distinta. Este workflow também combina bem com o Suno para adicionar trilha sonora ao vídeo final.

<!-- Case 25: K-Pop Choreography with Detailed Control (by @Kashberg_0) -->
### Caso 25: [Coreografia K-Pop — Controle detalhado](https://x.com/Kashberg_0/status/2049839091899088948) (por [@Kashberg_0](https://x.com/Kashberg_0))

Controle máximo sobre animação de dança: escreva uma coreografia de 16 passos com descrições precisas de movimento, gere a grade de referência com o GPT Image 2, depois anime com o Seedance 2.0. Cada passo recebe 2–3 segundos, produzindo um vídeo de dança contínuo de 35–50 segundos com qualidade autêntica de movimento K-pop.

<table><tr>

<td align="center"><video src="https://github.com/user-attachments/assets/1c088b5e-6305-4bf6-9377-97784d5f8fac" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Escreva uma sequência coreográfica detalhada (16 passos com movimentos de dança específicos)
2. Gere uma grade de referência mostrando cada passo com o GPT Image 2
3. Alimente a grade + a descrição completa da coreografia no Seedance 2.0
4. O modelo segue a sequência de passos com transições suaves


**Prompt para Seedance 2.0:**

```
K-Pop Dance Sequence (16 Steps, Korean Street)
[PROJECT TYPE]
Cinematic K-pop dance video (instruction-to-performance translation)
[CORE REQUIREMENT — STRICT]
The video must faithfully follow the exact 16-step choreography shown in the reference sheet, in the same order, with accurate poses and transitions.
No steps added, removed, or rearranged.
🧍‍♀️ [CHARACTER]
Korean female dancer (K-pop idol aesthetic)
Slim, athletic build
Same consistent face and proportions throughout
Expressive, confident stage presence
Natural, fluid but sharp K-pop movement quality
👕 [WARDROBE — K-POP STYLE]
Fitted crop top
Loose high-waisted jeans
Sneakers
Modern idol styling (clean, trendy)
Fabric reacts naturally to movement (denim weight, subtle folds)
📍 [LOCATION / ENVIRONMENT]
Empty aesthetic Korean street (Seoul-inspired)
Clean urban design: narrow street, minimal signage, soft architecture
No people, no vehicles
Slight cinematic depth (buildings, street lights, textures)
Lighting:
Soft daylight or golden hour (ideal for K-pop vibe)
Balanced highlights + gentle shadows
🔢 [16-STEP CHOREOGRAPHY — LOCKED SEQUENCE]
Starting Pose
Step Touch Right
Step Touch Left
Hip Sway Combo
Body Roll Down
Back Step Sweep
Quarter Turn Pivot
Hair Flip & Pose
Side Step Drag
Cross Behind Unwind
Body Wave Up
Hip Circle
Step Lock Step
Arm Sweep Pose
Chest Pop & Hit
Final Pose (hold 2–3 sec)
🎥 [CAMERA DIRECTION]
Full-body framing at all times
Start: centered wide shot
Smooth tracking + subtle dolly movement
Slight angle variation (front → 3/4 → side for spins)
No fast cuts — continuous flow
Camera movement complements choreography, not distracts
💃 [MOVEMENT STYLE — IMPORTANT]
Authentic K-pop choreography feel
Mix of:
Sharp hits (chest pop, accents)
Smooth transitions (body waves, turns)
Clean isolations (hips, chest, arms)
Controlled spins, balanced footwork
No jitter, no unnatural speed
⏱️ [TIMING]
Each step: ~2–3 seconds
Total duration: ~35–50 seconds
Seamless transitions between steps
🎵 [MUSIC DIRECTION — VERY IMPORTANT]
Genre: K-pop / K-pop instrumental / dance-pop
Tempo: 100–115 BPM
Style:
Clean beat drops
Punchy percussion
Light synth melodies
Modern idol choreography vibe
Sync Notes:
Step transitions hit beats
Step 8 (Hair Flip) hits a musical accent
Step 15 (Chest Pop) synced with a strong beat hit
Final pose lands on a clean musical ending
🎨 [VISUAL STYLE]
Photorealistic
Slightly stylized K-pop MV tone
Soft cinematic grading
Clean, polished, high-end look
⚙️ [OUTPUT SETTINGS]
4K resolution
24–30 FPS
High motion clarity
No distortion, no artifacts
🚫 [RESTRICTIONS]
No extra dancers
No background crowd
No outfit changes
No deviation from choreography
No camera cuts that break continuity
```

> [!NOTE]
> Quanto mais específicas forem suas descrições de passos, melhor o Seedance segue a coreografia. Nomeie movimentos de dança reais (body roll, hair flip, chest pop) em vez de descrições vagas. Esta técnica também funciona para kata de artes marciais, sequências de yoga ou qualquer movimento sequencial.

<!-- Case 27: Character Intro Animation (by @0xbisc) -->
### Caso 27: [Animação de introdução de personagem](https://x.com/0xbisc/status/2049496584283656690) (por [@0xbisc](https://x.com/0xbisc))

Crie uma animação de introdução de personagem em estilo de jogo AAA cyberpunk. Pegue qualquer imagem de personagem, redesenhe como personagem de jogo com o GPT Image 2, gere uma tela de introdução cinematográfica e depois anime a revelação com o Seedance 2.0. Substitua qualquer personagem — o workflow é agnóstico de personagem. 55 curtidas / 3K visualizações.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/intro_case27/output.jpg" width="400" alt="Saída de animação de introdução de personagem"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/e52eaa0b-b2fa-4c35-b790-a92af05d0c82" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Comece com uma imagem de personagem (sua própria arte, foto ou gerada por IA)
2. Use o GPT Image 2 para redesenhar como personagem de jogo AAA cyberpunk — mantenha a identidade facial, atualize o estilo
3. Gere uma tela de introdução cinematográfica com o personagem (fundo escuro, iluminação dramática, layout de cartão de título)
4. Anime a revelação de introdução no Seedance 2.0

**Prompt para GPT Image 2 — Redesign de personagem:**

```
based on the provided image, redesign as a cyberpunk AAA game character, keep face identity, keep original outfit, hyper-realistic game character, near-photoreal but still game-rendered, cinematic realism, in-game cutscene quality, cinematic lighting, strong contrast, realistic materials, depth of field, subject in sharp focus, background slightly blurred, strong foreground-background separation, Night City inspired environment, dense futuristic megacity, neon signage, wet streets, reflections, industrial details, fully human appearance, clean natural skin, no mechanical lines, no implants, no cyber patterns, character holding a highly designed futuristic weapon, dynamic action-ready pose, confident and intense expression, 16:9 AAA key visual, strong composition, character dominant, no logo, generate a unique character name fitting the character personality, character name in graffiti-style typography, medium-to-small size, integrated into layout, not dominant, refined character info module, editorial layout style, minimal, no background panel, only 1–2 short traits, extremely concise labels, grid-aligned typography-driven layout, Cyberpunk style UI, neon yellow text only, flat geometric layout, strict alignment, only one info module, no additional graphics, clean image, no heavy grain, no film grain, smooth surfaces, high polish, no anime, illustration, raw photography, metallic UI, gold color, cluttered layout, dense UI, boxes, background panels, color blocks, arrows, mechanical skin lines, cyber patterns

```

**Prompt para Seedance 2.0:**

```
industrial cyberpunk city at night, wet reflective ground, neon lights, distant explosions, floating sparks, cinematic atmosphere
camera always follows the character closely, no cuts, smooth tracking
motion continuity, no pose popping, no animation snapping, physically coherent transitions
0–2s:
character transitions into a low sliding movement
one hand brushing the ground for balance
sparks and debris react dynamically
weapon rotates forward in a smooth, deliberate motion
brief partial slow motion to emphasize control and flow
2–5s:
character raises weapon and fires while still moving forward
stylized compressed slow motion:
muzzle flash expands in layered light
face and muscles illuminated
subtle controlled recoil
shell casings eject in short slow-motion beats
particles and light distort around the shot
eyes focused strictly on target direction
final precise shot lands near the end of this phase
strong forward impact implied (sparks / explosion burst)
5–7s:
character motion fully stops, body settles naturally into final stance
character remains still, only subtle breathing motion
character lifts head and turns toward camera for the first time, then holds eye contact steadily
camera performs a subtle push-in
UI takes full visual focus:
UI builds progressively over the entire duration:
light glitch and scan effects
elements align into a clean layout
character name appears in graffiti handwritten animation, stroke-by-stroke reveal
secondary UI fades and slides in smoothly
```

> [!NOTE]
> Este workflow é agnóstico de personagem — substitua qualquer personagem (anime, realista, estilizado) e o pipeline se adapta. A chave é o processo em duas etapas do GPT Image 2: primeiro redesenhe o personagem para o estilo alvo, depois componha o layout da tela de introdução.

## 🎵 Videoclipe & Curta-metragem

<!-- Case 7: Music Video with Suno (by @fukaborichannel) -->
### Caso 7: [Videoclipe com Suno](https://x.com/fukaborichannel/status/2047206670020055317) (por [@fukaborichannel](https://x.com/fukaborichannel))

Combinação de três ferramentas: GPT Image 2 para visuais, Seedance 2.0 para movimento, Suno para música. Produza a música primeiro para travar o tempo e a estrutura, depois projete storyboards alinhados à batida.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/music_case7/input.jpg" width="280" alt="Storyboard gerado pelo GPT Image 2 para MV"></a></td>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/music_case7/output.jpg" width="280" alt="Quadro de saída do videoclipe"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/fd4be5c7-cd02-4a77-ae07-6b80efeff201" width="280" controls></video></td>
</tr></table>

**Passos:**

1. Gere a música no estilo desejado no Suno — confirme a estrutura da música (intro / verso / refrão)
2. Projete painéis de storyboard por seção da música no GPT Image 2
3. Anime cada painel no Seedance 2.0 — combine a duração do clipe com a batida
4. Sincronize os clipes com a faixa musical no seu software de edição


> [!NOTE]
> Produza a música primeiro. Conhecer o tempo e a duração antes de projetar o storyboard permite combinar precisamente o timing dos painéis com os cortes da batida.

<!-- Case 8: Cyberpunk Style Short Film (by @ponyodong) -->
### Caso 8: [Curta em estilo cyberpunk](https://x.com/ponyodong/status/2047210987263230133) (por [@ponyodong](https://x.com/ponyodong))

Use o GPT Image 2 para estabelecer um estilo visual consistente (cyberpunk, neon, lanternas, estética feminina), depois anime cada imagem com o Seedance 2.0 para produzir um curta estilizado que fica entre wallpaper, pôster e abertura de história.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/cyberpunk_case8/input.jpg" width="280" alt="Ilustração cyberpunk gerada pelo GPT Image 2"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/db6ebb63-90dc-47c5-96c5-ab2fa53ed56d" width="280" controls></video></td>
</tr></table>

**Passos:**

1. Defina o sistema de estilo visual no GPT Image 2 — fixe cores, iluminação e aparência do personagem
2. Gere de 4 a 6 imagens que carreguem o mesmo clima
3. Anime cada imagem no Seedance 2.0 com prompts de movimento lento e atmosférico
4. Sequencie os clipes para construir uma narrativa visual curta

<!-- Case 11: Japanese MV Full Toolchain (by @Tz_2022) -->
### Caso 11: [MV japonês — Toolchain completo de IA](https://x.com/Tz_2022/status/2047684399404056609) (por [@Tz_2022](https://x.com/Tz_2022))

Pipeline de quatro ferramentas produzindo um videoclipe completo em estilo japonês: GPT Image 2 para visuais → Seedance 2.0 para movimento → Suno 5.5 para música → CapCut para edição final. 742 curtidas / 107K visualizações.

<table><tr>
<td align="center"><video src="https://github.com/user-attachments/assets/e5ce621c-7fa3-47b5-99a7-00df7741a651" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Gere a música no Suno 5.5 primeiro — trave a duração, o tempo e o clima da música
2. Projete painéis de storyboard no GPT Image 2 cronometrados com as seções da música
3. Anime cada painel no Seedance 2.0, combinando a duração do clipe com a batida
4. Importe os clipes de vídeo e a faixa do Suno no CapCut — sincronize e exporte


> [!NOTE]
> Produza a música primeiro — conhecer a estrutura da batida antes de projetar os storyboards permite combinar precisamente o timing dos painéis com os cortes da música. Isso estende o Caso 7 (MV City Pop) adicionando o Suno ao loop e tratando todo o pipeline como uma produção sincronizada em vez de montagem posterior.

<!-- Case 20: Claude Shotlist → MV (by @CoffeeVectors) -->
### Caso 20: [Shotlist Claude → Videoclipe](https://x.com/CoffeeVectors/status/2049592150581485757) (por [@CoffeeVectors](https://x.com/CoffeeVectors))

Use o Claude para gerar uma shotlist detalhada (15 clipes de um segundo com diferentes ângulos de câmera e ações), gere um único retrato com o GPT Image 2, depois produza cada plano com o Seedance 2.0. Edite os clipes juntos com sua própria música do Suno para um MV completo. A IA escreve a direção criativa — você apenas executa.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/shotlist_case20/input.jpg" width="400" alt="Saída de videoclipe com shotlist Claude"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/d6ba86c4-65c3-4b1d-aa3c-846667f53b5e" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Gere um único retrato de personagem com o GPT Image 2 como âncora visual
2. Peça ao Claude para escrever uma shotlist de 15 planos (um plano por segundo) com ângulos e ações variados
3. Alimente o retrato + cada descrição de plano no Seedance 2.0 separadamente
4. Edite todos os clipes juntos e sincronize com sua faixa musical


**Prompt para Seedance 2.0 (por plano):**

```
A 15-second prestige-TV sequence, one shot per second, scored to an apocalyptic sacred crescendo — low organ and dissonant brass through roaring choir, hammered bells, and earth-shaking timpani to a final shattering harmonic strike. Throughout: a pale young queen with white hair, a tall ornate gold filigree crown, a translucent gauze veil, and a heavily jeweled pale gown — channeler of divine fire from above. Shot entirely handheld — visible micro-shake, breath-rhythm sway, reactive whip-corrections to action, documentary-tense framing.
1 (0–1s) Sky Opens. Handheld wide low-angle, camera tilted up. Black clouds spiral and split in a tear of white-gold light. She stands small below. Organ slams.
2 (1–2s) Eyes To Heaven. Handheld tight close-up, slight float. Her eyes lifted, gold light on her face, a tear of fire tracking down her cheek. Choir enters.
3 (2–3s) Hand Raised. Handheld medium, slight push-in. She raises one palm to the sky. Clouds above twist toward her gesture. Strings climb.
4 (3–4s) First Bolt. Handheld wide. A colossal pillar of holy fire descends and splits a distant black tower. Camera jolts on impact. Hammer beat.
5 (4–5s) The Pointing. Handheld tight medium. She extends one ringed finger slowly toward the horizon. Camera barely breathing. Bells ring.
6 (5–6s) Bolts Rain. Handheld wide, panning to track strikes. Dozens of pillars of holy fire descend across a battlefield. Camera whips reactively to each impact. Drums hammer.
7 (6–7s) Cloaked In Light. Handheld low-angle medium. A shaft of holy fire engulfs her without burning. Camera trembles in the pressure wave. Choir doubles.
8 (7–8s) The Wicked Burn. Handheld tight medium. A robed figure raises a blade — consumed in white-gold fire from above, ash silhouette collapsing. Camera flinches with the strike. Bass hit.
9 (8–9s) Walking Forward. Handheld tracking wide, operator moving with her. She advances across cracked scorched earth, pillars of fire descending in her wake. Strings shriek.
10 (9–10s) Crown Of Lightning. Handheld tight on the crown, slight float. White-gold lightning arcs continuously between the spires. Hair lifts in charged air. Bells climb.
11 (10–11s) Closed Fist. Handheld tight close-up. Her hand closes slowly into a fist. Vast clap of thunder. Camera shakes hard. Sustained held chord.
12 (11–12s) The Cleansing. Handheld wide, operator on a high vantage with visible sway. A fortified city struck by a grid of descending holy fire pillars. She stands small below, untouched. Choir at full roar.
13 (12–13s) The Quiet After. Handheld medium, breathing slowly. She lowers her hand. The storm stills. Ash falls like snow around her. Music drops to near-silence.
14 (13–14s) Eyes Return. Handheld extreme close-up, slight float. Eyes still warm gold blink once slowly. Faintest exhale. Single sustained tone.
15 (14–15s) The Smiting. Handheld frontal wide at dusk, settling into stillness on the final hold. She stands at the center of a vast scorched circle, horizon reduced to smoking ruin. Torn sky still glowing above her. Final shattering harmonic strike sustains.

Style: Photorealistic dark holy fantasy, prestige-TV aesthetic. Anamorphic 35mm, shallow DoF, heavy volumetric atmosphere — smoke, ash, ember haze, heat distortion, charged air shimmer. Palette of scorched bone-white, ivory, ash-gray, storm-slate, and incandescent white-gold. Painterly compositions, fine detail against destruction, organic film grain, heavy highlight bloom on the divine fire. Handheld throughout — visible micro-shake, reactive whip-corrections, breath-rhythm sway, camera flinching with every impact. No tripod stillness until the final hold. Operatic, terrifying, sovereign. The sky itself as her instrument.
```

> [!NOTE]
> Este workflow separa a direção criativa (Claude) da execução visual (GPT Image 2 + Seedance). É particularmente eficaz para videoclipes onde você precisa de muitos planos variados do mesmo personagem. O retrato único como âncora mantém a consistência em todos os 15 clipes.

## 🎮 Conceito de Jogo

<!-- Case 9: Game & Interactive Content (by @AbleGPT) -->
### Caso 9: [Jogos e conteúdo interativo](https://x.com/op7418/status/2046854932620525750) (por [@op7418](https://x.com/op7418))

Use o GPT Image 2 para gerar imagens de interface estilo jogo (com elementos de HUD, barras de habilidade, sobreposições de escolha), depois anime-as no Seedance 2.0 para simular sequências interativas de jogo. Estilos de jogo e ilustração enfrentam menos restrições de conteúdo no Seedance do que filmagens realistas de humanos.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/game_case9/input.jpg" width="400" alt="Imagem de entrada de interface de jogo"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/3d5d7525-b469-4c3b-aab9-68dc47630fdd" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Gere imagens estilo ARPG ou interface de jogo com o GPT Image 2, incluindo elementos de HUD
2. Importe no Seedance 2.0 e descreva a interação ou sequência de combate
3. Adicione efeitos de pós-produção (partículas, brilho) para polimento

**Prompt para GPT Image 2-2:**

```
帮我生成一个以《金瓶梅》为主题的古代 ARPG MMO 开放世界游戏的截图
```
**Prompt para GPT Image 2-2:**
```
出现 UI 选择 UI 之后变成第二张图的场景图
```

**Prompt para Seedance 2.0:**

```
选择 UI 之后变成第二张图右边的场景
```

**Variante — Simulação de jogo ARPG (por [@0xbisc](https://x.com/0xbisc/status/2047315350862352715)):**

One Piece, Stranger Things, qualquer IP — gere um screenshot de jogo de um mundo que não existe, depois expanda em gameplay ao vivo com o Seedance 2.0. 934 curtidas / 125K visualizações.

<table><tr>
<td align="center"><video src="https://github.com/user-attachments/assets/983b433a-88ea-4843-9047-fc01396752fe" width="400" controls></video></td>
</tr></table>

**Prompt para GPT Image 2:**

```
Generate an ARPG dialogue game screenshot inspired by [film/series name]
```

**Seedance 2.0:** Use o modo Image-to-Video. Nenhum prompt necessário — o Seedance lê o layout do HUD e o estende em uma sequência de gameplay automaticamente.

> [!NOTE]
> O Seedance 2.0 tem restrições sobre conteúdo humano realista. Estilos de jogo, anime e ilustração contornam a maioria dessas limitações e oferecem mais liberdade criativa.

<!-- Case 17: Game Interface Animation Full Pipeline (by @0xInk_) -->
### Caso 17: [Animação de interface de jogo — Pipeline completo](https://x.com/0xInk_/status/2048809000121360649) (por [@0xInk_](https://x.com/0xInk_))

O workflow mais viral desta coleção: crie uma animação completa de interface de videogame do zero. Comece com um personagem 2D no Midjourney, converta para visual 3D pronto para jogo com o GPT Image 2, projete a interface completa do jogo (HUD, telas de carregamento, menus), depois anime tudo com o Seedance 2.0. O GPT Image 2 se destaca aqui porque lida com detalhes de interface e permite retrabalho iterativo sem perda de qualidade. 2280 curtidas / 208K visualizações / 2793 salvamentos.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/game_case17/output.jpg" width="400" alt="Saída de animação de interface de jogo"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/b83da8f3-3dd6-44a3-bb27-b0d59cab381a" width="400" controls></video></td>
</tr></table>


> [!NOTE]
> O insight principal: o GPT Image 2 permite retrabalhar uma imagem várias vezes sem degradação de qualidade — perfeito para iterar em layouts de interface. Construa a interface completa do jogo como uma série de telas estáticas, depois deixe o Seedance conectá-las em uma animação fluida.

<!-- Case 24: GTA-Style City Game Concept (by @markgadala) -->
### Caso 24: [Conceito de jogo de cidade estilo GTA](https://x.com/markgadala/status/2048560337960489385) (por [@markgadala](https://x.com/markgadala))

Crie qualquer versão de GTA que quiser em 5 minutos. Gere screenshots de interface de jogo ambientados em qualquer cidade (Tóquio, Lagos, Mumbai) com o GPT Image 2, depois anime em filmagem de gameplay com o Seedance 2.0. O resultado parece um trailer real de um jogo que não existe. 99 curtidas / 8.7K visualizações.

<table><tr>
<td align="center"><video src="https://github.com/user-attachments/assets/d3b0a7b9-827a-47f6-b24e-eabfacf3e892" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Defina sua variante de GTA — cidade, época, estilo visual
2. Gere screenshots de jogo com o GPT Image 2: visão em terceira pessoa, sobreposição de HUD, ambiente urbano
3. Importe no Seedance 2.0 e anime como filmagem de gameplay
4. Monte os clipes em um trailer


> [!NOTE]
> Isso estende a abordagem de conceito de jogo do Caso 9 para jogos de mundo aberto urbano especificamente. Os elementos de HUD (minimapa, barra de vida, estrelas de procurado) são o que vendem a ilusão de "jogo real". Funciona para qualquer cidade — quanto mais específicos forem os detalhes ao nível da rua, mais convincente será o resultado.

## 🛠 Ferramentas de Produção

<!-- Case 18: Single Agent Automated Workflow (by @venturetwins) -->
### Caso 18: [Workflow automatizado com agente único](https://x.com/venturetwins/status/2048526911056613586) (por [@venturetwins](https://x.com/venturetwins))

A abordagem de esforço zero: diga a um único agente de IA (como o Glif) o que você quer, e ele cuida de todo o pipeline — gerando o storyboard com o GPT Image 2 e animando com o Seedance 2.0 — em uma única conversa. Sem transferência manual de arquivos, sem engenharia de prompt por etapa. 934 curtidas / 70K visualizações.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/agent_case18/output.jpg" width="400" alt="Saída do workflow automatizado com agente único"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/cc01849d-ee9b-47af-a7b0-d13250a001e0" width="400" controls></video></td>
</tr></table>

<!-- Case 21: Casting Grid Actor Audition (by @8fstudioz) -->
### Caso 21: [Grade de casting — Audição de atores](https://x.com/8fstudioz/status/2049547426198151627) (por [@8fstudioz](https://x.com/8fstudioz))

Economize créditos fazendo audição de 4 atores em uma única geração. Gere uma grade de casting com 4 painéis no GPT Image 2 mostrando diferentes atores para o mesmo papel, depois teste cada um no Seedance 2.0 com a mesma fala de diálogo. Descubra qual ator vale a pena investir mais créditos antes de se comprometer com um vídeo completo.

<table><tr>
<td align="center"><a href="https://evolink.ai/gpt-image-2-prompts?utm_source=github&utm_medium=picture&utm_campaign=gptimage2-x-seedance2"><img src="images/casting_case21/input.jpg" width="400" alt="Entrada da grade de casting"></a></td>
<td align="center"><video src="https://github.com/user-attachments/assets/dcdd958f-70cd-43f6-b191-4e0715fe2472" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Gere uma grade de casting com 4 painéis no GPT Image 2 — mesmo papel, 4 atores diferentes
2. Teste cada ator individualmente no Seedance 2.0 com o mesmo diálogo e ação
3. Compare a qualidade da performance (contato visual, expressão, movimento)
4. Invista os créditos restantes apenas no ator vencedor

**Prompt para GPT Image 2:**

```
Create a 16:9 horizontal cinematic casting board showing 4 different actor candidates for the same role.

Style:
[INSERT VISUAL STYLE]
Examples: CGI AAA video game cinematic, photorealistic, anime, stylized 3D

Role brief:
[INSERT ROLE DESCRIPTION]
Describe the type of lead or character the user is casting for.

World / genre:
[INSERT WORLD OR GENRE]
Examples: spy-action thriller, fantasy RPG, sci-fi adventure, crime drama

Wardrobe:
[INSERT WARDROBE DESCRIPTION]
Describe the clothing or outfit direction all 4 actors should share.

Tone:
[INSERT TONE]
Examples: sleek, dangerous, adventurous, grounded, moody, confident

Visual direction:
[INSERT VISUAL RENDERING NOTES]
Describe the rendering quality, material detail, realism level, facial detail, costume detail, and overall look.

Cinematic look:
[INSERT CINEMATIC STYLE]
Examples: blockbuster trailer aesthetic, prestige drama look, AAA game cinematic look

Camera framing:
[INSERT FRAMING]
Examples: 3/4 body, full body, waist-up

Camera angle:
[INSERT CAMERA ANGLE]
Examples: eye-level, slight low angle, slight 3/4 angle

Lens:
[INSERT LENS]
Examples: 50mm cinematic lens, 85mm portrait lens

Depth of field:
[INSERT DEPTH OF FIELD]
Examples: shallow, shallow but controlled

Lighting:
[INSERT LIGHTING SETUP]
Describe the lighting style.

Background:
[INSERT BACKGROUND DESCRIPTION]
Describe the background environment or backdrop.

Colour treatment:
[INSERT COLOUR TREATMENT]
Describe the grading or colour tone.

Layout:
Arrange the 4 actor candidates in a 16:9 horizontal composition with 4 evenly spaced vertical panels across the frame, one actor per panel from left to right.

Character variation:
Each candidate should feel like a different casting choice for the same role. Vary facial structure, age feel, hairstyle, expression, posture, and energy, but keep them grounded in the same world, wardrobe logic, and tonal universe.

Important:
- Same role
- Same world
- Same wardrobe logic
- Same visual style
- Different actor interpretations
- No duplicated faces
- No text
- No labels
- No watermark

The final image should feel like a premium cinematic casting board for [INSERT PROJECT TYPE].
Examples: a film, a game, an animated short, a cinematic trailer
```

**Prompt para Seedance 2.0 (por ator):**

```
Use the uploaded 16:9 four-panel casting board as the source image.

Create a controlled 15-second cinematic casting audition reel for [INSERT ROLE OR PROJECT TYPE].

Animate the actors one by one in this exact order from left to right:

0.0–3.5 seconds: ONLY the far-left actor performs.
3.5–7.0 seconds: ONLY the second actor from the left performs.
7.0–10.5 seconds: ONLY the third actor from the left performs.
10.5–14.0 seconds: ONLY the far-right actor performs.
14.0–15.0 seconds: hold on the full four-panel board with all actors still.

Each actor delivers the same audition line:
"[INSERT DIALOGUE LINE]"

Performance direction:
Each actor should look directly into the camera while delivering the line, as if performing a screen test audition. Their eye line should stay locked to camera.

Each actor should deliver the line with:
[INSERT PERFORMANCE TRAITS]
Examples: calm control, quiet menace, emotional vulnerability, confidence, charm, intensity, humor

The performance should feel:
[INSERT PERFORMANCE TONE]
Examples: sleek, cinematic, believable, grounded, dramatic, stylized

Each actor should bring a slightly different interpretation of the same role.

Control rules:
ONLY the active actor moves during their assigned time window.
ONLY the active actor speaks during their assigned time window.
ONLY animate the active actor's mouth, eyes, facial expression, head, and subtle upper-body movement.
The active actor must look directly at the camera while speaking.
All other actors remain completely still like frozen reference images.
Do not animate multiple actors at the same time.
Do not change the panel layout.
Do not change actor positions.
Do not cut to a new scene.
Do not reframe into a different composition.
Do not change wardrobe.
Do not change background.
Do not change lighting.
Do not add new characters.
Do not add extra dialogue.
Do not add captions, subtitles, labels, or text.

Camera direction:
Keep the four-panel 16:9 casting board as the main composition. Use only [INSERT CAMERA MOVEMENT STYLE] toward the active actor during their performance window.
Examples: a subtle cinematic push-in, gentle focus emphasis, minimal controlled emphasis

Keep the movement [INSERT CAMERA BEHAVIOUR].
Examples: minimal, smooth, controlled

Keep the actor presented toward camera so the audition feels direct and comparable.

Audio / timing:
Each actor should speak the dialogue clearly within about 3.5 seconds.
The same line is repeated four times, once per actor.
No overlapping voices.
No background conversation.
No unnecessary sound effects.

Final result:
A clean casting audition reel where four actor candidates perform the same line one by one from left to right, each looking directly into the camera, making it easy to compare screen presence, facial acting, eye contact, posture, and dialogue delivery.
```

> [!NOTE]
> Um personagem pode ficar ótimo em uma imagem estática, mas perder completamente o papel quando você testa diálogo, contato visual e performance. Este workflow antecipa a decisão de casting antes de gastar créditos em cenas completas.

<!-- Case 22: 3D Sculpt → AI Render → Animation (by @_DAntunes_) -->
### Caso 22: [Escultura 3D → Render IA → Animação](https://x.com/_DAntunes_/status/2049142166232904078) (por [@_DAntunes_](https://x.com/_DAntunes_))

Conecte modelagem 3D tradicional com vídeo de IA: crie um modelo 3D bruto em argila no Nomad Sculpt (ou qualquer app de escultura), use o GPT Image 2 para renderizá-lo em uma ilustração polida, depois anime com o Seedance 2.0 via ComfyUI. Isso dá controle preciso sobre pose e composição que prompts de texto puro não conseguem alcançar.

<table><tr>
<td align="center"><video src="https://github.com/user-attachments/assets/f5ecdb0c-d1ca-4291-91bc-eb88de91cd82" width="400" controls></video></td>
</tr></table>

**Passos:**

1. Esculpa um modelo 3D bruto no Nomad Sculpt (ou Blender, ZBrush, etc.)
2. Exporte um screenshot do modelo no ângulo de câmera desejado
3. Use o GPT Image 2 para renderizar o modelo 3D em uma ilustração polida ou imagem realista
4. Importe a imagem renderizada no Seedance 2.0 (via ComfyUI ou direto) e anime

> [!NOTE]
> O modelo 3D oferece algo que nenhum prompt de texto consegue: controle exato sobre pose corporal, posição das mãos e ângulo de câmera. Mesmo um modelo bruto em argila é suficiente — o GPT Image 2 cuida de toda a renderização e detalhamento. Este pipeline é ideal para criadores que já usam ferramentas 3D e querem adicionar animação de IA ao seu workflow.

## 🌟 Vitrine da Comunidade

Uma vitrine contínua de trabalhos com **GPT Image 2 × Seedance 2.0** compartilhados por criadores no X. Clique em qualquer vídeo para reproduzir; clique no nome do autor para abrir o post original. As adições mais recentes aparecem primeiro.

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/cd87bc68-5613-410e-8dbe-df656aaa504d" width="240" controls></video>
<br/><a href="https://x.com/abxxai/status/2055636095736709190"><b>@abxxai</b></a> · <sub>80,365 views</sub>
<br/><sub>You can literally create anything with AI right now.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/fd7dac8a-0984-416d-8632-72ecd3106586" width="240" controls></video>
<br/><a href="https://x.com/pabloprompt/status/2055726656287871478"><b>@pabloprompt</b></a> · <sub>75,035 views</sub>
<br/><sub>One Piece BTS · PART 2 😍</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6ae8904b-6510-4aa5-ba8f-12ab001669ec" width="240" controls></video>
<br/><a href="https://x.com/maarcoofdezz/status/2057371189207584943"><b>@maarcoofdezz</b></a> · <sub>30,023 views</sub>
<br/><sub>Just made a full cinematic ad for a luxury mens cologne usi…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/484aa607-4d00-4543-95d0-7838fe7ce0db" width="240" controls></video>
<br/><a href="https://x.com/AIwithAliya/status/2055674114845925710"><b>@AIwithAliya</b></a> · <sub>24,084 views</sub>
<br/><sub>This looks so neat and clean</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/0c2f7cf0-f7e1-4f52-b7b7-e53c179fcc0d" width="240" controls></video>
<br/><a href="https://x.com/HeyOz_AI/status/2057871032480825398"><b>@HeyOz_AI</b></a> · <sub>17,459 views</sub>
<br/><sub>Claude + GPT Image 2 + seedance + Meta ads MCP Replaced my…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/77e02dda-8b51-4654-8af2-1bbb93f822d6" width="240" controls></video>
<br/><a href="https://x.com/Creatify_AI/status/2056431801971999147"><b>@Creatify_AI</b></a> · <sub>16,350 views</sub>
<br/><sub>Creatify AI agents use the BEST AI models for each process.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/3adf863a-165b-4d3e-bb20-e427cd3ce648" width="240" controls></video>
<br/><a href="https://x.com/MayorKingAI/status/2057933848948965741"><b>@MayorKingAI</b></a> · <sub>14,113 views</sub>
<br/><sub>I created a steampunk action sequence in a hand-painted 3D…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/800f2bbc-d53b-4075-8e08-b63a2e69c3e3" width="240" controls></video>
<br/><a href="https://x.com/IATheYoker/status/2057402859222933891"><b>@IATheYoker</b></a> · <sub>12,503 views</sub>
<br/><sub>GPT Image 2 + Seedance 2.5 ya pueden crear intros del Mundi…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/008d22d8-9a47-4051-ab60-7743958cc4c8" width="240" controls></video>
<br/><a href="https://x.com/ShamiWeb3/status/2055832098435629466"><b>@ShamiWeb3</b></a> · <sub>9,644 views</sub>
<br/><sub>A TAPNOW luxury skincare visual in which a glowing beauty m…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/592c1944-d3bf-4db6-b546-f152a836081d" width="240" controls></video>
<br/><a href="https://x.com/GlitterPixely/status/2057903405712953505"><b>@GlitterPixely</b></a> · <sub>8,862 views</sub>
<br/><sub>Quick character highlight intro with GPT Image 2 + Seedance…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/4467f8e3-192a-4937-95a6-1945122a6d66" width="240" controls></video>
<br/><a href="https://x.com/luche_whitewing/status/2055626188241150350"><b>@luche_whitewing</b></a> · <sub>7,837 views</sub>
<br/><sub>深夜の美女👠 lovart (@lovart_jp)× 🎬 GPT-image-2 × Seedance 2.0 で作…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/4d6c8938-e4a2-4bf6-ab16-e7a2ed4d8205" width="240" controls></video>
<br/><a href="https://x.com/DoctorAmna11/status/2057108821537861739"><b>@DoctorAmna11</b></a> · <sub>6,903 views</sub>
<br/><sub>The Ghost Who Couldn’t Scare Anyone</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ec26a6f4-f6ca-4a4b-bc68-3c11d818982d" width="240" controls></video>
<br/><a href="https://x.com/I_amShiti/status/2057378173856494070"><b>@I_amShiti</b></a> · <sub>5,830 views</sub>
<br/><sub>One AI app gives you access to👇</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/f7323047-0d46-4cba-945f-e8199a837787" width="240" controls></video>
<br/><a href="https://x.com/eijo_AIart/status/2056935787581898972"><b>@eijo_AIart</b></a> · <sub>5,028 views</sub>
<br/><sub>PolloAI GPT-Image2 x Seedance2で、映画製作用ストーリーボードを使った、ショートフィルム制…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/01aa8296-2c8c-4c99-a915-2bbdc9232125" width="240" controls></video>
<br/><a href="https://x.com/aziz4ai/status/2057824424838004836"><b>@aziz4ai</b></a> · <sub>4,885 views</sub>
<br/><sub>JADE FALCON: AWAKENING</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/70eea8b8-d3fa-4555-9c60-52ed8bed4711" width="240" controls></video>
<br/><a href="https://x.com/opener_ai/status/2055627299429777684"><b>@opener_ai</b></a> · <sub>4,353 views</sub>
<br/><sub>Gpt image 2 x Seedance 2.0 @dreamina_ai</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/014d3a33-b68a-4977-8f72-a239b65a835c" width="240" controls></video>
<br/><a href="https://x.com/spect3ral/status/2057871834310148242"><b>@spect3ral</b></a> · <sub>4,319 views</sub>
<br/><sub>Claude + GPT Image 2 + Seedance + Meta Ads MCP</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/2f8d430d-fd0f-4d1b-9c74-92f85fa466b2" width="240" controls></video>
<br/><a href="https://x.com/mdmadeit/status/2055685500279697883"><b>@mdmadeit</b></a> · <sub>3,970 views</sub>
<br/><sub>made a full 1:25 anime short</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ea9bbdfd-a5b1-4e85-a886-362362f6e3cb" width="240" controls></video>
<br/><a href="https://x.com/Strength04_X/status/2055849751317524982"><b>@Strength04_X</b></a> · <sub>3,737 views</sub>
<br/><sub>Created this cinematic sushi 🍣storyboard with GPT Image 2 +…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/595b4af8-ece2-4cca-98df-982721e685bd" width="240" controls></video>
<br/><a href="https://x.com/jasminekhan90_/status/2057399215228322261"><b>@jasminekhan90_</b></a> · <sub>3,418 views</sub>
<br/><sub>Human connection still matters in an AI future.</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6e646b59-7485-470d-af4e-de983dda6b12" width="240" controls></video>
<br/><a href="https://x.com/Sogni_Protocol/status/2057822871817191549"><b>@Sogni_Protocol</b></a> · <sub>3,342 views</sub>
<br/><sub>There’s no better combo on the market right now than GPT Im…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/28bb17d7-6fc0-4f94-8716-8b5871a290ea" width="240" controls></video>
<br/><a href="https://x.com/noorlewisx/status/2055863201389240369"><b>@noorlewisx</b></a> · <sub>2,974 views</sub>
<br/><sub>Made with Seedance 2 + GPT Image 2</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/b265ac7b-fb01-4e4c-954c-f4ab748f7fc4" width="240" controls></video>
<br/><a href="https://x.com/markgadala/status/2055896597662228650"><b>@markgadala</b></a> · <sub>2,876 views</sub>
<br/><sub>AI makes creating Pixar quality animations so incredibly ea…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/f4c4152c-15d0-404b-8040-f86d54ecbc28" width="240" controls></video>
<br/><a href="https://x.com/SocialSight/status/2057334346223177973"><b>@SocialSight</b></a> · <sub>2,720 views</sub>
<br/><sub>did you get the news?</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/4ba27b9b-0431-419e-90fe-6a25816979c5" width="240" controls></video>
<br/><a href="https://x.com/hanifproduktif/status/2055828657172820269"><b>@hanifproduktif</b></a> · <sub>2,595 views</sub>
<br/><sub>Replaced (GPT Image 2 + Seedance 2.5)</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/11375c25-a923-44a1-a8a4-8d04acecab31" width="240" controls></video>
<br/><a href="https://x.com/adrianaia_/status/2057474538594607520"><b>@adrianaia_</b></a> · <sub>2,382 views</sub>
<br/><sub>De las calles frías a un salón de clases lleno de amor.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/0dd27e90-469f-4ee8-aebb-3f3afbf2d8f2" width="240" controls></video>
<br/><a href="https://x.com/itsPolloAI/status/2057333580133724593"><b>@itsPolloAI</b></a> · <sub>2,153 views</sub>
<br/><sub>🎉 Pollo AI × GPT Image 2 × Seedance 2.0 — Results are in.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/c62c2125-a2c4-4e3a-9e99-e6f94124086c" width="240" controls></video>
<br/><a href="https://x.com/SimplyAnnisa/status/2058068924806160785"><b>@SimplyAnnisa</b></a> · <sub>2,026 views</sub>
<br/><sub>Golden mornings and buttery saltbread bliss</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/7e1d91f7-0153-4b5d-ac13-ae62da10f517" width="240" controls></video>
<br/><a href="https://x.com/Artedeingenio/status/2057401307510481397"><b>@Artedeingenio</b></a> · <sub>1,479 views</sub>
<br/><sub>Using Niji to create children’s illustrations, GPT Image 2…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/5a72d57d-f9fb-42a4-8e15-2de535d3b6a8" width="240" controls></video>
<br/><a href="https://x.com/hasamaru_studio/status/2057433716339933656"><b>@hasamaru_studio</b></a> · <sub>1,461 views</sub>
<br/><sub>GPT Image 2 でショットリストを作成し、 Seedance 2.0 のリファレンス生成で動画を作成しました。</sub>
</td>
<td align="center" valign="top" width="25%"></td>
<td align="center" valign="top" width="25%"></td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/35c556af-1752-4a8a-9965-d1e2314b242e" width="240" controls></video>
<br/><a href="https://x.com/Pixelbunny_ai/status/2051985506414768154"><b>@Pixelbunny_ai</b></a>
<br/><sub>- Create Stunning AAA quality shorts with leading models -…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/75051d48-988e-4e0d-9cfa-389d821abec8" width="240" controls></video>
<br/><a href="https://x.com/Adam38363368936/status/2051969842748735596"><b>@Adam38363368936</b></a>
<br/><sub>GPT image 2+Seedance 2</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/8241abd7-3b4e-4b62-8dca-24dac31926cc" width="240" controls></video>
<br/><a href="https://x.com/ai_hakase_/status/2051950389063282894"><b>@ai_hakase_</b></a>
<br/><sub>【AIでNetflix級のUIを爆速生成！GPT Image 2 × Seedance 2.0】 👉   最新のAIを…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/30710ea7-1f06-49b2-ad31-041fa95046c2" width="240" controls></video>
<br/><a href="https://x.com/Hoshimiko_AIart/status/2051947486353433013"><b>@Hoshimiko_AIart</b></a>
<br/><sub>「見たいアニメに間に合わない……！！」</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/8315ea16-32ff-4078-8dce-44d4d2c896d9" width="240" controls></video>
<br/><a href="https://x.com/Dheer_Red/status/2051915196185346333"><b>@Dheer_Red</b></a>
<br/><sub>Seedance 2.0, Veo 3.1, Nano Banana, GPT Image 2—all in one…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/c57c7caf-ab27-4c79-9ca6-b571d7899139" width="240" controls></video>
<br/><a href="https://x.com/NyaiiBubu/status/2051914243193389078"><b>@NyaiiBubu</b></a>
<br/><sub>AI for UGC modal Rp0 itu nyata 😭</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/1899a3a5-da5c-455b-8fa3-d3c3ab830df4" width="240" controls></video>
<br/><a href="https://x.com/apoorvabr/status/2051904397324722349"><b>@apoorvabr</b></a>
<br/><sub>I like the video concepts of @chrisfirst.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/1718996f-d0e8-4ba8-abef-1d0e725ed22f" width="240" controls></video>
<br/><a href="https://x.com/MaAyyoub/status/2051900019444154376"><b>@MaAyyoub</b></a>
<br/><sub>Don't ruin a new day by thinking about yesterday.</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/a7d5b81e-708e-4bb5-9263-7bfb9c6e0d01" width="240" controls></video>
<br/><a href="https://x.com/halne369/status/2051867032333803722"><b>@halne369</b></a>
<br/><sub>Seedance2.0用の絵コンテ作成のスキルができました！</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/4a76e9c0-0171-4b2a-a74d-90f3e59f80e8" width="240" controls></video>
<br/><a href="https://x.com/RYD232210555420/status/2051847984053469424"><b>@RYD232210555420</b></a>
<br/><sub>朋友设计的无人驾驶公交车 Sumgo，我帮它做了条 AI 概念宣传片。</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ac2e370f-6986-4ac1-b2ec-9580901d9483" width="240" controls></video>
<br/><a href="https://x.com/Lart_AI/status/2051838590016241772"><b>@Lart_AI</b></a>
<br/><sub>🎮 Built with GPT Image 2 × Seedance 2.0 on LartAI!</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ad9b0494-b689-4aad-8a50-bcd278110c8b" width="240" controls></video>
<br/><a href="https://x.com/Jake_Joseph/status/2051774091108155844"><b>@Jake_Joseph</b></a>
<br/><sub>Wait, you can put real screenshots inside AI-generated UGC…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/661be30d-946d-42f6-888c-fc19d8ab6e9d" width="240" controls></video>
<br/><a href="https://x.com/KimAkiyama81/status/2051768139566714958"><b>@KimAkiyama81</b></a>
<br/><sub>Choreographing a hallway action scene using GPT Image 2 and…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6735dd89-bc7d-462e-8f5f-69fab82158e7" width="240" controls></video>
<br/><a href="https://x.com/ChangningL29508/status/2051743657980748119"><b>@ChangningL29508</b></a>
<br/><sub>Use the same storyboard to generate a realistic character r…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/d3d201fc-0813-4d08-aab0-eb0d7a27f708" width="240" controls></video>
<br/><a href="https://x.com/aiseomastery/status/2051733667106734129"><b>@aiseomastery</b></a>
<br/><sub>THIS AI WORKFLOW TURNS A RAMEN RECIPE INTO A STUNNING ANIME…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ae6c44a0-8e6a-4ff1-ab1f-99b7aa49af01" width="240" controls></video>
<br/><a href="https://x.com/simplissimus_/status/2051714965485039897"><b>@simplissimus_</b></a>
<br/><sub>Quer dominar a Força e ver sua própria versão Jedi ganhar v…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/67479af0-384f-471f-a8ef-4ee4727289f0" width="240" controls></video>
<br/><a href="https://x.com/azed_ai/status/2051693299376021888"><b>@azed_ai</b></a>
<br/><sub>Studios sell this as pre-production</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/1800bbc0-e6f2-4289-9d41-ef07c15be380" width="240" controls></video>
<br/><a href="https://x.com/fy360593/status/2051686764054790504"><b>@fy360593</b></a>
<br/><sub>Been seeing a lot of people doing this "Fan cam" content la…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/954dc98f-ab00-4fd5-9106-8671310958ef" width="240" controls></video>
<br/><a href="https://x.com/indrawan_ape/status/2051680370429685963"><b>@indrawan_ape</b></a>
<br/><sub>GPT Image 2 × Seedance 2.0 on @higgsfield is insane.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/fbf50a8f-459b-47a9-8ab1-3cc89b91d239" width="240" controls></video>
<br/><a href="https://x.com/floopers966/status/2051678374750203983"><b>@floopers966</b></a>
<br/><sub>また最安値入札ミサイルかよ！</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/c793c91a-d280-4c2d-b1fd-d64271f467a8" width="240" controls></video>
<br/><a href="https://x.com/Xaroon_x/status/2051656676441293172"><b>@Xaroon_x</b></a>
<br/><sub>Made with GPT Image 2 + Seedance 2.5 by @yapper_so</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/998a71a7-0481-4e93-a717-36295181449c" width="240" controls></video>
<br/><a href="https://x.com/MonetizationDon/status/2051644080803750092"><b>@MonetizationDon</b></a>
<br/><sub>I decided to create my own Afrobeats Mortal Kombat-style sh…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/eaadef33-8db9-4aad-9c2e-a455fcf896ad" width="240" controls></video>
<br/><a href="https://x.com/roomA708/status/2051639024574697952"><b>@roomA708</b></a>
<br/><sub>GPT-Image-2 × Seedance 2.0で、Osmo Pocket 4の“架空CM”を作ってみました。</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6a66e25d-2ae1-469f-896c-3f0593a79f8f" width="240" controls></video>
<br/><a href="https://x.com/vkuoo/status/2051637837951615142"><b>@vkuoo</b></a>
<br/><sub>Using Midjourney to generate the original images, GPT Image…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/b6fdbff6-23c8-4c6e-9642-7c7fca112b35" width="240" controls></video>
<br/><a href="https://x.com/RadineerE10/status/2051622937808318654"><b>@RadineerE10</b></a>
<br/><sub>「YouMind」が世界最大級の無料AIプロンプトライブラリとして存在感を増している。</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/0fc459c5-03bc-4fce-928c-40ff34acf989" width="240" controls></video>
<br/><a href="https://x.com/aivoxyy/status/2051621547518083130"><b>@aivoxyy</b></a>
<br/><sub>GPT Image 2 + Seedance 2.5 a police chase of new 2026 Chevr…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ec1de9ac-97a4-4648-bab6-5ae8818377c1" width="240" controls></video>
<br/><a href="https://x.com/aadi29494/status/2051594382437232667"><b>@aadi29494</b></a>
<br/><sub>Made a LEGO build-process video with GPT Image 2 + Seedance…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/cdb64846-fbfc-4b0f-8a18-9a55552eff87" width="240" controls></video>
<br/><a href="https://x.com/mmarch_ai/status/2051591272918299085"><b>@mmarch_ai</b></a>
<br/><sub>Con GPT Image 2 dominas la composición: añade textos largos…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/c2ca0649-d276-4ae6-9d1f-e68a0907fc20" width="240" controls></video>
<br/><a href="https://x.com/D_studioproject/status/2051580845606191260"><b>@D_studioproject</b></a>
<br/><sub>How to join a cult with GPT Image 2 x Seedance 2.0 Anime St…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6f206801-4d40-44a0-971b-3500f7619f1b" width="240" controls></video>
<br/><a href="https://x.com/EgeUymaz/status/2051576423089901874"><b>@EgeUymaz</b></a>
<br/><sub>Storyboarded with GPT Image 2.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/399714ae-3ac1-4540-b1db-be9320051ee0" width="240" controls></video>
<br/><a href="https://x.com/CurieuxExplorer/status/2051536554691334385"><b>@CurieuxExplorer</b></a>
<br/><sub>Quiet Growth 🌱</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/da772e93-8226-4ac6-aff8-6b9786c861ce" width="240" controls></video>
<br/><a href="https://x.com/josepho/status/2051535229161021618"><b>@josepho</b></a>
<br/><sub>My new AI minishort, Dance of destruction, based on an old…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/acf55ef2-63e7-420a-9d9a-61b97f2165f9" width="240" controls></video>
<br/><a href="https://x.com/iswangwenbin/status/2051528434225234302"><b>@iswangwenbin</b></a>
<br/><sub>我也来交作业了👇 Hyperframes + Mimo TTS + GPT Image 2 + Seedance 2.5</sub>
</td>
<td align="center" valign="top" width="25%"></td>
<td align="center" valign="top" width="25%"></td>
<td align="center" valign="top" width="25%"></td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/4fbcf8e4-c509-4ddd-8961-855881b3c2c9" width="240" controls></video>
<br/><a href="https://x.com/HAL2400AI/status/2050076981702906004"><b>@HAL2400AI</b></a> · <sub>6,721,336 views</sub>
<br/><sub>ドンキで爆速品出しするゲームのプレイ映像。</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/66649bde-1a17-4ea7-92b9-1b06b5e6a5d8" width="240" controls></video>
<br/><a href="https://x.com/0xbisc/status/2050154597340287143"><b>@0xbisc</b></a> · <sub>224,490 views</sub>
<br/><sub>Kitchen Hunt</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/33ef696a-45ff-4a35-84d3-4ee0c65629cc" width="240" controls></video>
<br/><a href="https://x.com/ElevenCreative/status/2049866735898009836"><b>@ElevenCreative</b></a> · <sub>115,974 views</sub>
<br/><sub>Create UGC videos in minutes with ElevenCreative.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/3c851335-abd7-46bb-9f44-7e5b4ce44d8b" width="240" controls></video>
<br/><a href="https://x.com/Saccc_c/status/2049769037660360897"><b>@Saccc_c</b></a> · <sub>115,759 views</sub>
<br/><sub>用 GPT Image 2 + Seedance 2.5，还原了故宫太和殿的建造全过程🤩</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/d4902050-6912-4ab1-b204-cb373e0eafe3" width="240" controls></video>
<br/><a href="https://x.com/ZaraIrahh/status/2049668274330255388"><b>@ZaraIrahh</b></a> · <sub>55,203 views</sub>
<br/><sub>Created with Gpt Image 2 + Seedance 2.0 on @SJinn_Agent</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/d710c64b-8b05-4848-ae11-16a5ac303416" width="240" controls></video>
<br/><a href="https://x.com/johnAGI168/status/2050190239398781120"><b>@johnAGI168</b></a> · <sub>49,957 views</sub>
<br/><sub>The future of AI video is here, and it is completely mind-b…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/8da15a4b-4f5e-47c7-a97b-891ef8227b52" width="240" controls></video>
<br/><a href="https://x.com/GumVue/status/2050314912094687425"><b>@GumVue</b></a> · <sub>45,606 views</sub>
<br/><sub>Create with a Custom GPT  - Short Film Prompt Generator :…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/b78cdad1-5f5c-4309-880d-6efdadf0a18d" width="240" controls></video>
<br/><a href="https://x.com/seiiiiiiiiiiru/status/2049983823308570686"><b>@seiiiiiiiiiiru</b></a> · <sub>41,147 views</sub>
<br/><sub>Midjourney V8.1 ↓ GPT image 2.0 ↓ Seedance 2.0</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6f3b85bc-b409-4c1a-8fdb-325d9d6b0291" width="240" controls></video>
<br/><a href="https://x.com/VORTEX_Promos/status/2049614941812863376"><b>@VORTEX_Promos</b></a> · <sub>40,410 views</sub>
<br/><sub>TOP 7 INSANE GPT Image 2 x Seedance 2.0 Prompts You Must Try</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/3cb1548c-59a8-4e80-8969-759e7f6ff7d9" width="240" controls></video>
<br/><a href="https://x.com/ElCopyMaster/status/2049513235121144002"><b>@ElCopyMaster</b></a> · <sub>39,205 views</sub>
<br/><sub>Pollo AI acaba de cambiar la creación de anuncios 🤯</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/e4dbbfac-0b0c-4fed-a9f3-309da015cc0e" width="240" controls></video>
<br/><a href="https://x.com/4111J_/status/2049817671714443561"><b>@4111J_</b></a> · <sub>31,198 views</sub>
<br/><sub>What’s in my bag?</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/a29e9b73-e176-4e46-9377-22bf5c2d4f89" width="240" controls></video>
<br/><a href="https://x.com/AIwithkhan/status/2049732042695623134"><b>@AIwithkhan</b></a> · <sub>28,965 views</sub>
<br/><sub>Here we go GPT Image 2 and Seedance 2.0 is now live on @ins…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/3e29d94c-a175-4df9-8568-4b0707c5f775" width="240" controls></video>
<br/><a href="https://x.com/rovvmut_/status/2049900959028109567"><b>@rovvmut_</b></a> · <sub>28,036 views</sub>
<br/><sub>GPT Image 2 and Seedance 2.0 on @insmind_com</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/c8919582-b859-4b38-b456-a02683700f1e" width="240" controls></video>
<br/><a href="https://x.com/aimikoda/status/2049589670581874777"><b>@aimikoda</b></a> · <sub>27,837 views</sub>
<br/><sub>Gpt Image 2 + Seedance 2.0 Trailer Workflow</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/aed7449c-40b7-4160-8a6f-2a824b0d95f2" width="240" controls></video>
<br/><a href="https://x.com/HBCoop_/status/2049853870172410026"><b>@HBCoop_</b></a> · <sub>20,844 views</sub>
<br/><sub>Decided to test myself out in the storyboard workflow.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/028cd6eb-d4cd-4ab3-8484-4d895173d8a6" width="240" controls></video>
<br/><a href="https://x.com/franpradasAI/status/2050168636321440010"><b>@franpradasAI</b></a> · <sub>19,487 views</sub>
<br/><sub>🚨 NOVEDAD: Un anuncio así cuesta $2M</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6fd777ca-b6d9-43b9-a4a4-3b2b41a4fd04" width="240" controls></video>
<br/><a href="https://x.com/john_my07/status/2049524601471074422"><b>@john_my07</b></a> · <sub>19,165 views</sub>
<br/><sub>Crafted this one by turning a movement sheet as a reference…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/31aef043-4277-4f3e-bcca-d79d6747f9c1" width="240" controls></video>
<br/><a href="https://x.com/MrLarus/status/2050505429529104752"><b>@MrLarus</b></a> · <sub>18,511 views</sub>
<br/><sub>太飒了！ 🤯用ChatGPT+Seedance生成街头篮球1v1，真实感很强！  运球试探、crossover变向、欧…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/8fad0bd0-251f-4602-9573-4d2b2bdd2c53" width="240" controls></video>
<br/><a href="https://x.com/LaTwitchance/status/2050183976657047764"><b>@LaTwitchance</b></a> · <sub>17,313 views</sub>
<br/><sub>Une vidéo de gameplay montre un jeu où il faut réapprovisio…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/741ee092-27aa-40b9-959c-0ec5e2944a62" width="240" controls></video>
<br/><a href="https://x.com/Marco_Exito/status/2050219814678098028"><b>@Marco_Exito</b></a> · <sub>14,156 views</sub>
<br/><sub>💥ÚLTIMA HORA:  ¡GPT-IMAGE-2 y Seedance 2.0 ya están disponi…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/a410a383-9744-4f1f-8ba6-92eba2cdbec4" width="240" controls></video>
<br/><a href="https://x.com/JoyLi629/status/2049740242085667272"><b>@JoyLi629</b></a> · <sub>13,056 views</sub>
<br/><sub>GPT image2 + Seedance 2.0 做产品爆炸视频💥太香了</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/a6b9b482-32fa-41e7-bfb2-f97b2cf46722" width="240" controls></video>
<br/><a href="https://x.com/doctorwasif/status/2050244639253569886"><b>@doctorwasif</b></a> · <sub>12,897 views</sub>
<br/><sub>Chatgpt GPT-2 & Seedance 2 on @yapper_so</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/c00ed740-192d-4d5b-a981-a6664c9bce07" width="240" controls></video>
<br/><a href="https://x.com/code_bykuti/status/2049852139112182181"><b>@code_bykuti</b></a> · <sub>10,592 views</sub>
<br/><sub>We’re not generating images anymore…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/af52163e-dab5-4c14-8531-c3e9f4fcf2d8" width="240" controls></video>
<br/><a href="https://x.com/igus_ai/status/2049527468143329284"><b>@igus_ai</b></a> · <sub>9,010 views</sub>
<br/><sub>Ahora puedes crear wallpapers de cualquier jugador o equipo…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/98001464-cbf5-4dd8-8161-a8d7a86d0afd" width="240" controls></video>
<br/><a href="https://x.com/Diplomeme/status/2049752171474993460"><b>@Diplomeme</b></a> · <sub>8,011 views</sub>
<br/><sub>Makeup tutorial (ft.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/e14db781-59cf-4ad2-8cbc-41fa99bbbebc" width="240" controls></video>
<br/><a href="https://x.com/XMonetizationC_/status/2049527585269551150"><b>@XMonetizationC_</b></a> · <sub>7,509 views</sub>
<br/><sub>GRWM (ft.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/0b0e902b-325e-42e5-98f1-57177f73124e" width="240" controls></video>
<br/><a href="https://x.com/nrqa__/status/2049836461755949357"><b>@nrqa__</b></a> · <sub>7,319 views</sub>
<br/><sub>GPT-IMAGE-2 &amp; Seedance 2.0 is now officially available…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/acba4f2e-b8aa-47ea-8a08-e05d44523d45" width="240" controls></video>
<br/><a href="https://x.com/IqraSaifiii/status/2049845664880955862"><b>@IqraSaifiii</b></a> · <sub>7,004 views</sub>
<br/><sub>Create your own Influencer live stream 🔥</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/f952740e-5a89-442e-a45a-38fd496acdef" width="240" controls></video>
<br/><a href="https://x.com/IntLab0000/status/2050227050502639625"><b>@IntLab0000</b></a> · <sub>6,444 views</sub>
<br/><sub>【Seedance 2.0でSora2を目指すテスト】Seedanceに直接長文のプロンプトで依頼する代わりに、gpt…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ef66fa7a-c595-4b42-88e0-bd81464caa6b" width="240" controls></video>
<br/><a href="https://x.com/weiinberg/status/2049867302309707927"><b>@weiinberg</b></a> · <sub>6,126 views</sub>
<br/><sub>GPT Image 2 x Seedance 2.0 on @insmind_com</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/c21ebea6-629e-494c-b8ea-b8fb5b6b2d87" width="240" controls></video>
<br/><a href="https://x.com/angeldot_/status/2049528144256737579"><b>@angeldot_</b></a> · <sub>6,114 views</sub>
<br/><sub>Puedes crear wallpapers animados como este en segundos</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/9c0fcc5d-b76e-4b6d-b123-9d65837b1263" width="240" controls></video>
<br/><a href="https://x.com/Just_sharon7/status/2050430486548447502"><b>@Just_sharon7</b></a> · <sub>6,084 views</sub>
<br/><sub>Are you ready to buy this necklace??</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/f4171c4c-923a-4af1-aeb2-11ec47174f17" width="240" controls></video>
<br/><a href="https://x.com/Sheldon056/status/2049832802078838857"><b>@Sheldon056</b></a> · <sub>5,910 views</sub>
<br/><sub>GPT Image 2 and Seedance 2.0 are now live on @insmind_com</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/bc924494-934c-4175-a32d-8b8ef66fc07e" width="240" controls></video>
<br/><a href="https://x.com/nicos_ai/status/2049526583212327013"><b>@nicos_ai</b></a> · <sub>5,776 views</sub>
<br/><sub>Ahora puedes crear Wallpapers animados en segundos</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/1e2860b3-802e-4d5d-953d-a81c1e5e8d03" width="240" controls></video>
<br/><a href="https://x.com/promptsref/status/2050489527475507706"><b>@promptsref</b></a> · <sub>5,635 views</sub>
<br/><sub>Create a short film like this in just 1 minute with GPT Ima…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/7205e175-7e35-44a2-9b3e-3e0d65e0ca72" width="240" controls></video>
<br/><a href="https://x.com/AIwithSynthia/status/2049818475846332898"><b>@AIwithSynthia</b></a> · <sub>5,248 views</sub>
<br/><sub>Excited to inform you that GPT Image 2 and Seedance 2.0 are…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/b6a422ee-1bf8-4098-b463-b393c2d449f5" width="240" controls></video>
<br/><a href="https://x.com/wanerfu/status/2050457318194815166"><b>@wanerfu</b></a> · <sub>5,165 views</sub>
<br/><sub>我用动作参考图来制作舞蹈动画，使用了 Seedance 2.0 + GPT Image 2.0</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/dcec0956-895f-4ab8-80e8-889f4efb2e7f" width="240" controls></video>
<br/><a href="https://x.com/FinanceYF5/status/2049672873888116956"><b>@FinanceYF5</b></a> · <sub>4,039 views</sub>
<br/><sub>有人用 GPT Image 2 × Seedance 2.0 在 OpenArt 上制作一支万宝路广告。</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/31ffd0f8-273c-460b-acd3-3e3fea5ef3fc" width="240" controls></video>
<br/><a href="https://x.com/meng_dagg695/status/2049739564630307176"><b>@meng_dagg695</b></a> · <sub>3,885 views</sub>
<br/><sub>GPT image 2 × Seedance 2.0 combo 🔥 on @yapper_so</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/9d4a97cf-c95b-495b-bf62-34b8b95c6d4b" width="240" controls></video>
<br/><a href="https://x.com/AI_Arabic1/status/2049534174646940051"><b>@AI_Arabic1</b></a> · <sub>3,867 views</sub>
<br/><sub>شوفوا الإبداع 😱!!</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/fc83dd63-3809-47e9-9578-c97f235327f2" width="240" controls></video>
<br/><a href="https://x.com/AIwithSarah_/status/2049838433582067838"><b>@AIwithSarah_</b></a> · <sub>3,566 views</sub>
<br/><sub>GPT Image 2 and Seedance 2.0 is now available on @insmind_c…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/af16374f-c77b-4929-9ede-9ea6f15551ed" width="240" controls></video>
<br/><a href="https://x.com/MatiasSchrank/status/2050185710561276382"><b>@MatiasSchrank</b></a> · <sub>3,564 views</sub>
<br/><sub>Así lo hice con Smart Shot de @openart_ai:</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/04043531-d981-4798-ade7-87e1bcd70144" width="240" controls></video>
<br/><a href="https://x.com/ivnways/status/2050149994691523033"><b>@ivnways</b></a> · <sub>3,303 views</sub>
<br/><sub>ÚLTIMA HORA: ¡GPT-IMAGE-2 y Seedance 2.0 ya disponibles gra…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/a7734353-2509-4b77-96cc-6b42e6fd4699" width="240" controls></video>
<br/><a href="https://x.com/SwayamShrma/status/2049970629450088960"><b>@SwayamShrma</b></a> · <sub>2,908 views</sub>
<br/><sub>From rough Idea to short film just by using a workflow that…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/fd1436bc-ebaf-4706-82d7-7ee4e5056963" width="240" controls></video>
<br/><a href="https://x.com/Alex_Inspira/status/2050210855242084796"><b>@Alex_Inspira</b></a> · <sub>2,888 views</sub>
<br/><sub>Así es exactamente como lo hice.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/7f246962-7346-4e52-9049-6f7cc31e1f3f" width="240" controls></video>
<br/><a href="https://x.com/oggii_0/status/2049859002398556580"><b>@oggii_0</b></a> · <sub>2,655 views</sub>
<br/><sub>I tested a full AI food commercial workflow using ONE promp…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/3dcc955f-6441-4c57-bf77-371c5d51ebb7" width="240" controls></video>
<br/><a href="https://x.com/Ciri_ai/status/2049861486575747436"><b>@Ciri_ai</b></a> · <sub>2,603 views</sub>
<br/><sub>One prompt → full cinematic sequence all generated inside @…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/fa0834a6-a585-492a-b92f-22ed8e9d155b" width="240" controls></video>
<br/><a href="https://x.com/Taaruk_/status/2049702491768684839"><b>@Taaruk_</b></a> · <sub>2,485 views</sub>
<br/><sub>GPT IMAGE 2 + seedance 2.0</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/27c79a49-058b-45d9-ae76-8bb9d0612571" width="240" controls></video>
<br/><a href="https://x.com/noahsolomon/status/2050145785891914119"><b>@noahsolomon</b></a> · <sub>2,477 views</sub>
<br/><sub>I love using fal to make developing fal easier.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/043c67b9-c0a6-450e-a061-0173ddde350b" width="240" controls></video>
<br/><a href="https://x.com/harboriis/status/2050083212664426882"><b>@harboriis</b></a> · <sub>2,389 views</sub>
<br/><sub>A complete dance sequence mapped frame by frame</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/8f1842af-4ea2-46a5-9719-64aafc14d512" width="240" controls></video>
<br/><a href="https://x.com/techyoutbe/status/2049819747597029627"><b>@techyoutbe</b></a> · <sub>2,354 views</sub>
<br/><sub>One workspace.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/e6b408d1-888f-47d5-bb70-2aa4d041d666" width="240" controls></video>
<br/><a href="https://x.com/ZetoGroovin/status/2050176480861094147"><b>@ZetoGroovin</b></a> · <sub>2,298 views</sub>
<br/><sub>『神降ろし』- AI MUSIC VIDEO - 和の雰囲気の曲と映像でMV作ってみました。</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/dd6544dc-4107-4bd1-a7ea-35e6bf23f0a0" width="240" controls></video>
<br/><a href="https://x.com/0xtonixie/status/2050449563773981106"><b>@0xtonixie</b></a> · <sub>2,165 views</sub>
<br/><sub>GPT的images2 ＋seedance2.0绝对是制作AI漫剧的最佳组合：</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/49846a04-64a8-4bff-bac8-b193f62256d3" width="240" controls></video>
<br/><a href="https://x.com/iamrealsnow/status/2049770674886152232"><b>@iamrealsnow</b></a> · <sub>1,764 views</sub>
<br/><sub>GRWM (Male Minimal Editorial) Using GPT image 2.0 + Seedanc…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/26160c30-a15b-4de8-b946-df6573d70c0d" width="240" controls></video>
<br/><a href="https://x.com/budgetpixel/status/2049684059278676332"><b>@budgetpixel</b></a> · <sub>1,754 views</sub>
<br/><sub>GPT Image 2.0 + Seedance 2.0 = a powerful dance workflow</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/f070db63-ea05-41bb-8b55-9f697ab6844a" width="240" controls></video>
<br/><a href="https://x.com/0kncn/status/2050264466852368557"><b>@0kncn</b></a> · <sub>1,720 views</sub>
<br/><sub>Tesla vs Einstein Tekken evreninde</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/1ff92c6d-0131-4147-840b-82322df72bb7" width="240" controls></video>
<br/><a href="https://x.com/iX00AI/status/2050356236814885066"><b>@iX00AI</b></a> · <sub>1,673 views</sub>
<br/><sub>実際の流れ👇  1文入力 → Shot Plan（設計図）生成 → そのまま映像生成  プロンプトの工夫も編集も不要。…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/b57613c4-0838-410c-9a33-785b6002b34e" width="240" controls></video>
<br/><a href="https://x.com/ashen_one/status/2049601589573292154"><b>@ashen_one</b></a> · <sub>1,608 views</sub>
<br/><sub>GPT 2.0 + Seedance 2.0 = Brand Goldmine</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/92da6dee-b575-4ba9-a405-d94a3572b18f" width="240" controls></video>
<br/><a href="https://x.com/obrunookamoto/status/2050156722086461498"><b>@obrunookamoto</b></a> · <sub>1,338 views</sub>
<br/><sub>A Higgsfield conectou 30+ modelos de geração de imagem e ví…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/6a4f6292-f7fe-4a20-a6f1-4eee47e5ccad" width="240" controls></video>
<br/><a href="https://x.com/Kashberg_0/status/2050117126397329698"><b>@Kashberg_0</b></a> · <sub>1,291 views</sub>
<br/><sub>GPT IMAGE 2 and Seedance 2.0</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ed99f828-225c-4b28-a667-879290c4838a" width="240" controls></video>
<br/><a href="https://x.com/sara4ai/status/2050189053979496926"><b>@sara4ai</b></a> · <sub>1,229 views</sub>
<br/><sub>مرحبا يا أصدقاء  خلال هذا الأسبوع لاحظت أعمالًا مبهرة باستخ…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/9718f4a9-8466-4622-af0a-03788c2b27ae" width="240" controls></video>
<br/><a href="https://x.com/bmx_ai13/status/2050432594647642414"><b>@bmx_ai13</b></a> · <sub>1,220 views</sub>
<br/><sub>A super simple workflow: 2 character images → GPT Image 2.0…</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/14da241f-97fb-4162-ad1c-77e54ab3c29c" width="240" controls></video>
<br/><a href="https://x.com/ChillaiKalan__/status/2049833865825632522"><b>@ChillaiKalan__</b></a> · <sub>1,196 views</sub>
<br/><sub>GPT Image 2 and Seedance 2.0 on @insmind_com</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/ab068907-e9b7-4762-846e-2d0ca3adb48d" width="240" controls></video>
<br/><a href="https://x.com/suji_pop/status/2050135943294878035"><b>@suji_pop</b></a> · <sub>1,173 views</sub>
<br/><sub>Subway Surfer in a post apocalyptic world, with the graphic…</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/5207f874-d9a2-47c2-afe9-6c0b07b87e47" width="240" controls></video>
<br/><a href="https://x.com/densancar/status/2050210215237173629"><b>@densancar</b></a> · <sub>1,148 views</sub>
<br/><sub>Claude Code + GPT 2 + Seedance 2.0 Makes VIRAL AI UGC Videos</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/681faaa8-d52f-427d-b192-b4ac8b12ebc0" width="240" controls></video>
<br/><a href="https://x.com/nexudotio/status/2049775355586576758"><b>@nexudotio</b></a> · <sub>1,128 views</sub>
<br/><sub>Open Design now ships images AND videos.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/a9a2a2e0-ab3f-45e1-8eeb-903c2e72bccf" width="240" controls></video>
<br/><a href="https://x.com/OiiOii_AI/status/2049777567193043420"><b>@OiiOii_AI</b></a> · <sub>1,108 views</sub>
<br/><sub>OiiOiiで、話題のプロンプトを GPT Image 2 × Seedance 2.0 で試してみました。</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/70b7d00f-a84a-4b96-b53f-9ffc3cf25a67" width="240" controls></video>
<br/><a href="https://x.com/QingQ77/status/2050201770320949363"><b>@QingQ77</b></a> · <sub>1,051 views</sub>
<br/><sub>cool 使用 GPT Image 2 + Seedance 2.5 创建游戏界面的视频动画</sub>
</td>
</tr></table>

<table><tr>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/d9a22507-f0b6-4186-886c-33129a23d651" width="240" controls></video>
<br/><a href="https://x.com/Parul_Gautam7/status/2050184972942946328"><b>@Parul_Gautam7</b></a> · <sub>1,044 views</sub>
<br/><sub>Here’s how it actually works.</sub>
</td>
<td align="center" valign="top" width="25%">
<video src="https://github.com/user-attachments/assets/a77e53bb-5764-4fd7-b44e-82b369195a83" width="240" controls></video>
<br/><a href="https://x.com/joshesye/status/2049715911838355739"><b>@joshesye</b></a> · <sub>1,020 views</sub>
<br/><sub>茶叶品牌 TVC《一口春山》</sub>
</td>
<td align="center" valign="top" width="25%"></td>
<td align="center" valign="top" width="25%"></td>
</tr></table>

## 💡 Dicas & Técnicas

### Guia de consistência

Manter a consistência visual entre as saídas do GPT Image 2 e ao longo da animação do Seedance 2.0 é o desafio mais comum. Estas abordagens tratam cada camada.

**Consistência de imagem de produto**

A causa raiz da distorção de produto no Seedance é que sua interpolação de movimento reescreve detalhes finos — logos, texturas e padrões de superfície são modificados.

Soluções:
- Adicione `keep the product appearance completely unchanged, camera movement only, no rotation` ao seu prompt do Seedance
- Escolha movimento de câmera (push-in, pull-out) em vez de movimento do objeto — mantenha o produto parado e mova a câmera
- Mantenha a duração do clipe abaixo de 3 segundos — clipes mais curtos acumulam menos distorção

**Consistência de personagem**

- Gere uma folha de personagem com três vistas primeiro e use-a como âncora visual para todos os quadros de storyboard subsequentes
- Inclua uma breve descrição do personagem (cor do cabelo, roupa, porte) em cada prompt de painel de storyboard
- Evite trocar a perspectiva do personagem dentro de um único clipe

**Consistência de cena**

Ao gerar múltiplos painéis de storyboard no GPT Image 2, fixe os parâmetros de cena no topo do seu prompt:

```
Scene setting: [location], [time of day], [lighting direction], [fixed background elements].
Maintain this scene setting unchanged across all panels.
```

---

### Templates de prompt

**GPT Image 2 → Template de storyboard**

```
Create a [N]-panel storyboard for [subject]:
- Style: [realistic / anime / illustration / cinematic]
- Aspect ratio: 16:9 widescreen
- Each panel: include shot type (wide / medium / close-up) + action description
- Character: [fixed appearance description]
- Scene tone: [color palette / lighting / mood]
Output as a single image with [N] panels separated by thin lines.
```

**GPT Image 2 → Template de grade 3×3**

```
Output a single 3×3 grid storyboard image showing the following continuous action:
[describe the action sequence]
Requirements:
- 9 panels arranged left-to-right, top-to-bottom showing continuous motion
- Character position and scale consistent across all panels
- Background consistent throughout
- No text, labels, or content outside the panel borders
```

**Seedance 2.0 → Template estilo anime**

```
Japanese full-color animation, high-speed editing, high frame count, 24fps.
[Scene description]. [Character description]. [Action description].
Strong camera work, high visual impact.
```

**Seedance 2.0 → Template estilo comercial**

```
Cinematic commercial quality, [brand tone: premium / energetic / warm],
[product] centered in frame, slow camera push-in,
[lighting direction] highlights the product, clean background, no people.
Duration: 3 seconds.
```

**Tamanho do prompt — mais curto frequentemente vence**

Experimento da comunidade via [@Iancu_ai](https://x.com/Iancu_ai/status/2047882924679168083): um prompt de 1500 palavras com qualidade cinematográfica para o Seedance perdeu para uma única frase. Mesmo personagem, mesmos 15 segundos. O prompt curto venceu. O Seedance recompensa clareza direcional em vez de descrição exaustiva — escreva a intenção de movimento, não cada detalhe da cena.

## 🚀 Experimente no Evolink

O Evolink permite executar tanto o GPT Image 2 quanto o Seedance 2.0 em um só lugar — sem trocar de plataforma, sem re-upload de arquivos.

**Por que Evolink**

- Uma única chave de API para GPT Image 2 e Seedance 2.0
- Transferência direta de imagem para vídeo na mesma interface — gere uma imagem e clique em "Enviar para Vídeo" sem precisar baixar
- Processamento em lote — enfileire múltiplos painéis de storyboard para geração sequencial de vídeo

**Como usar**

```
Step 1: Open Evolink → select GPT Image 2 → generate your storyboard image
Step 2: Click "Generate Video" → Seedance 2.0 receives the image automatically
Step 3: Add your Seedance prompt → generate
```

<a href='https://evolink.ai/signup?utm_source=github&utm_medium=readme&utm_campaign=gptimage2-x-seedance2'><img src='https://img.shields.io/badge/🚀 Get%20Started-Evolink-black' height="25"></a>


## 🙏 Agradecimentos

Este repositório foi inspirado por coleções excepcionais de workflows abertos e experimentos compartilhados pela comunidade.

Agradecemos aos criadores e colaboradores que compartilharam seu trabalho publicamente e tornaram estes estudos de caso possíveis:
[@szounft](https://x.com/szounft) · [@Toshi_nyaruo_AI](https://x.com/Toshi_nyaruo_AI) · [@ponyodong](https://x.com/ponyodong) · [@servasyy_ai](https://x.com/servasyy_ai) · [@YaReYaRu30Life](https://x.com/YaReYaRu30Life) · [@fukaborichannel](https://x.com/fukaborichannel) · [@Shin_Engineer](https://x.com/Shin_Engineer) · [@ai_mitosan](https://x.com/ai_mitosan) · [@kiyoshi_shin](https://x.com/kiyoshi_shin) · [@AbleGPT](https://x.com/AbleGPT) · [@patata1216](https://x.com/patata1216) · [@peter6759](https://x.com/peter6759) · [@hibi_ai__](https://x.com/hibi_ai__) · [@heygentlewhale](https://x.com/heygentlewhale) · [@ai_gezgini](https://x.com/ai_gezgini) · [@Tz_2022](https://x.com/Tz_2022) · [@old_pgmrs_will](https://x.com/old_pgmrs_will) · [@0xbisc](https://x.com/0xbisc) · [@Iancu_ai](https://x.com/Iancu_ai) · [@Jake_Joseph](https://x.com/Jake_Joseph) · [@venturetwins](https://x.com/venturetwins) · [@0xInk_](https://x.com/0xInk_) · [@markgadala](https://x.com/markgadala) · [@Ankit_patel211](https://x.com/Ankit_patel211) · [@Ciri_ai](https://x.com/Ciri_ai) · [@nimentrix](https://x.com/nimentrix) · [@insmind_com](https://x.com/insmind_com) · [@kingofdairyque](https://x.com/kingofdairyque) · [@Kashberg_0](https://x.com/Kashberg_0) · [@airina_xyz](https://x.com/airina_xyz) · [@CoffeeVectors](https://x.com/CoffeeVectors) · [@mdmadeit](https://x.com/mdmadeit) · [@Morph_VGart](https://x.com/Morph_VGart) · [@MEnesKirca](https://x.com/MEnesKirca) · [@MrLarus](https://x.com/MrLarus) · [@AYi_AInotes](https://x.com/AYi_AInotes) · [@8fstudioz](https://x.com/8fstudioz) · [@_DAntunes_](https://x.com/_DAntunes_) · [@promptsref](https://x.com/promptsref) · [@Just_sharon7](https://x.com/Just_sharon7) · [@wanerfu](https://x.com/wanerfu) · [@AIwithkhan](https://x.com/AIwithkhan) · [@0xtonixie](https://x.com/0xtonixie) · [@doctorwasif](https://x.com/doctorwasif) · [@HAL2400AI](https://x.com/HAL2400AI) · [@bmx_ai13](https://x.com/bmx_ai13) · [@ZaraIrahh](https://x.com/ZaraIrahh) · [@iX00AI](https://x.com/iX00AI) · [@GumVue](https://x.com/GumVue) · [@adriansolarzz](https://x.com/adriansolarzz) · [@0kncn](https://x.com/0kncn) · [@john_my07](https://x.com/john_my07) · [@XMonetizationC_](https://x.com/XMonetizationC_) · [@harboriis](https://x.com/harboriis) · [@IntLab0000](https://x.com/IntLab0000) · [@Marco_Exito](https://x.com/Marco_Exito) · [@Alex_Inspira](https://x.com/Alex_Inspira) · [@densancar](https://x.com/densancar) · [@QingQ77](https://x.com/QingQ77) · [@johnAGI168](https://x.com/johnAGI168) · [@sara4ai](https://x.com/sara4ai) · [@MatiasSchrank](https://x.com/MatiasSchrank) · [@Parul_Gautam7](https://x.com/Parul_Gautam7) · [@LaTwitchance](https://x.com/LaTwitchance) · [@ZetoGroovin](https://x.com/ZetoGroovin) · [@franpradasAI](https://x.com/franpradasAI) · [@obrunookamoto](https://x.com/obrunookamoto) · [@ivnways](https://x.com/ivnways) · [@noahsolomon](https://x.com/noahsolomon) · [@OiiOii_AI](https://x.com/OiiOii_AI) · [@suji_pop](https://x.com/suji_pop) · [@SuguruKun_ai](https://x.com/SuguruKun_ai) · [@aimikoda](https://x.com/aimikoda) · [@seiiiiiiiiiiru](https://x.com/seiiiiiiiiiiru) · [@SwayamShrma](https://x.com/SwayamShrma) · [@IqraSaifiii](https://x.com/IqraSaifiii) · [@rovvmut_](https://x.com/rovvmut_) · [@ashen_one](https://x.com/ashen_one) · [@weiinberg](https://x.com/weiinberg) · [@ElevenCreative](https://x.com/ElevenCreative) · [@SunNeverSetsX](https://x.com/SunNeverSetsX) · [@oggii_0](https://x.com/oggii_0) · [@HBCoop_](https://x.com/HBCoop_) · [@code_bykuti](https://x.com/code_bykuti) · [@AIwithSarah_](https://x.com/AIwithSarah_) · [@nrqa__](https://x.com/nrqa__) · [@ChillaiKalan__](https://x.com/ChillaiKalan__) · [@Sheldon056](https://x.com/Sheldon056) · [@techyoutbe](https://x.com/techyoutbe) · [@AIwithSynthia](https://x.com/AIwithSynthia) · [@4111J_](https://x.com/4111J_) · [@hrrcnes](https://x.com/hrrcnes) · [@nexudotio](https://x.com/nexudotio) · [@iamrealsnow](https://x.com/iamrealsnow) · [@Saccc_c](https://x.com/Saccc_c) · [@Raul_IA_Prod](https://x.com/Raul_IA_Prod) · [@Diplomeme](https://x.com/Diplomeme) · [@JoyLi629](https://x.com/JoyLi629) · [@meng_dagg695](https://x.com/meng_dagg695) · [@bindureddy](https://x.com/bindureddy) · [@FinanceYF5](https://x.com/FinanceYF5) · [@joshesye](https://x.com/joshesye) · [@Taaruk_](https://x.com/Taaruk_) · [@budgetpixel](https://x.com/budgetpixel) · [@VORTEX_Promos](https://x.com/VORTEX_Promos) · [@AI_Arabic1](https://x.com/AI_Arabic1) · [@angeldot_](https://x.com/angeldot_) · [@igus_ai](https://x.com/igus_ai) · [@nicos_ai](https://x.com/nicos_ai) · [@ElCopyMaster](https://x.com/ElCopyMaster)

*Não podemos garantir que cada caso esteja atribuído ao criador original. Se algo precisar ser corrigido, entre em contato conosco e atualizaremos.*

Se você tem mais casos de workflow interessantes para compartilhar, sinta-se à vontade para entrar em contato e nos ajudar a expandir a biblioteca de workflows do Evolink.

[![Star History Chart](https://api.star-history.com/svg?repos=EvoLinkAI/GPT-Image-2-Seedance-2.5-Workflow&type=Date)](https://www.star-history.com/#EvoLinkAI/GPT-Image-2-Seedance-2.5-Workflow&Date)