awesome-seedance-2.5-prompts-skills
100+ curated Seedance 2.5 prompts with real video previews, plus an installable Agent Skill that optimizes prompts, creates storyboards, and generates videos via Seedance models.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Check it deeper on your machine
It has changed since its scan on 11 Nov 2026.
As a Claude Code plugin, install it and then ask Claude Code:
/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills with scanskillsafety
Or from a terminal, clone it and run it:
git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills'
It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.
Badge for your README
Markdown
[](/listings/atlascloudai--awesome-seedance-2-5-prompts-skills)HTML
<a href="/listings/atlascloudai--awesome-seedance-2-5-prompts-skills"><img src="/api/listings/atlascloudai--awesome-seedance-2-5-prompts-skills/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Install the workflow Skill, the Universal Video Prompt Skill it references, and the default Atlas Cloud execution Skill:
npx skills add AtlasCloudAI/awesome-seedance-2.5-prompts-skills --skill seedance-2-5-skill npx skills add AtlasCloudAI/awesome-seedance-2.5-prompts-skills --skill universal-video-prompt-skill npx skills add AtlasCloudAI/atlas-cloud-skills --skill atlas-cloud
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 2,563 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.
- # 🎬 Awesome Seedance 2.5 Prompts + Seedance 2.5 Skill
- [](https://awesome.re)
- [](https://creativecommons.org/licenses/by/4.0/)
- [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills)
- [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/pulls)
- [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills)
- A curated library of Seedance 2.5 video prompts plus **Seedance 2.5 Skill**, an installable Agent Skill that improves prompts, plans and reviews storyboards when needed, and generates controllable video. It uses Seedream 5.0 Pro for storyboard images and Seedance 2.0 as the current executable video default, switching to Seedance 2.5 only when the selected provider exposes it.
- | [Browse prompts](https://www.atlascloud.ai/prompts-hub/seedance-2-5-prompt?utm_source=github&utm_campaign=awesome-seedance-2.5-prompts-skills) | [Submit your prompt](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/issues/new?template=prompt.yml) | [Install Seedance 2.5 Skill](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/tree/main/skills/seedance-2-5-skill) | [Generate with Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_campaign=awesome-seedance-2.5-prompts-skills) | [Get an API key](https://www.atlascloud.ai/console/api-keys?utm_source=github&utm_campaign=awesome-seedance-2.5-prompts-skills) |
- |---|---|---|---|---|
- [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/README.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_zh.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_zh-TW.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_ja.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_ko.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_th.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_vi.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_hi.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_es.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_de.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_fr.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_it.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_pt.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_tr.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_ar.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_id.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_nl.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_pl.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_ru.md) [](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/i18n/README_sv.md)
- ---
- <a id="contents"></a>
- ## 📖 Contents
- - [🤔 What is Seedance 2.5?](#model-overview)
- - [🧩 Seedance 2.5 prompt guide](#prompt-guide)
- - [🧠 Seedance 2.5 Skill](#seedance-2-5-skill)
- - [🌐 Universal Video Prompt Skill](#universal-video-prompt-skill)
- - [🚀 How to use this repository](#how-to-use)
- - [⚙️ Model and execution defaults](#model-and-execution-defaults)
- - [🔎 Curation and provenance](#curation-and-provenance)
- - [📊 Statistics](#statistics)
- - [🔥 Featured Prompts](#featured-prompts)
- - [🏷️ Browse by Category](#browse-by-category)
- - [📋 All Prompts](#all-prompts)
- - [❓ Frequently asked questions](#faq)
- - [🔗 Resources](#resources)
- - [📄 License](#license)
- <a id="model-overview"></a>
- ## 🤔 What is Seedance 2.5?
- Seedance 2.5 is ByteDance's next-generation multimodal video generation model, following Seedance 2.0. It is designed for reference-driven video creation, longer narrative sequences, synchronized audio and video, and precise visual control.
- **Launch status:** Seedance 2.5 is expected to launch in August 2026. Atlas Cloud is one of the first official API launch partners for Seedance 2.5.
- Published Seedance 2.5 launch material describes up to 30-second generation, native 4K output, up to 50 multimodal references, and local region editing. Treat these as announced capabilities, not universal API parameters.
- **Availability note:** provider availability, reference limits, duration, resolution, and editing controls can differ. The Skill verifies actual model availability instead of assuming Seedance 2.5 is callable.
- <a id="prompt-guide"></a>
- ## 🧩 Seedance 2.5 prompt guide
- ### Prompt structure used by the Skill
- 1. **Reference binding:** state what each image, video, or audio reference controls.
- 2. **Observable action:** describe visible events in temporal order, including reactions and state changes.
- 3. **Spatial relationships:** say where subjects, objects, and the camera are in relation to each other.
- 4. **Camera and cuts:** specify framing, movement, cut order, match actions, and occlusions only where they matter.
- 5. **Visual style:** define lighting, palette, texture, atmosphere, and pace.
- 6. **Audio:** define dialogue, ambience, sound effects, or music when the selected model supports them.
- 7. **Constraints:** preserve only the identities, product details, scene traits, and exclusions that are essential.
- <a id="seedance-2-5-skill"></a>
- ## 🧠 Seedance 2.5 Skill
- **Seedance 2.5 Skill** is the name of the installable Agent Skill in this repository. It turns a creative brief, existing prompt, references, or a storyboard into a production-ready prompt, the necessary intermediate assets, an execution request, and a review loop.
- ### Install
- Install the workflow Skill, the Universal Video Prompt Skill it references, and the default Atlas Cloud execution Skill:
- ```bash
- npx skills add AtlasCloudAI/awesome-seedance-2.5-prompts-skills --skill seedance-2-5-skill
- npx skills add AtlasCloudAI/awesome-seedance-2.5-prompts-skills --skill universal-video-prompt-skill
- npx skills add AtlasCloudAI/atlas-cloud-skills --skill atlas-cloud
- ```
- ### What it does
- - **Improve prompts:** turn an idea, an existing prompt, or reference assets into a clearer production-ready Seedance prompt.
- - **Plan and review storyboards:** create a Seedream 5.0 Pro storyboard only when the video needs multi-shot planning or stronger visual consistency, display it as progress, and review its quality automatically.
- - **Generate and review video:** use the best available Seedance model, submit through the selected Atlas Cloud channel, poll the same task to completion, and check the result.
- <details><summary>Skill files and references</summary>
- - [`SKILL.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/SKILL.md)
- - [`references/workflow.zh-CN.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/workflow.zh-CN.md)
- - [`references/long-video.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/long-video.md)
- - [`references/multi-reference.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/multi-reference.md)
- - [`references/real-person.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/real-person.md)
- - [`references/transitions.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/transitions.md)
- - [`references/editing-and-extension.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/editing-and-extension.md)
- - [`references/capabilities.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/capabilities.md)
- - [`references/model-profile.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/model-profile.md)
- - [`references/cinematography.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/cinematography.md)
- - [`references/prompt-blocks.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/prompt-blocks.md)
- - [`references/prompt-templates.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/prompt-templates.md)
- - [`references/execution-adapters.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/execution-adapters.md)
- - [`references/troubleshooting.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/references/troubleshooting.md)
- - [`scripts/generate.mjs`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/seedance-2-5-skill/scripts/generate.mjs)
- </details>
- <a id="universal-video-prompt-skill"></a>
- ## 🌐 Universal Video Prompt Skill
- **Universal Video Prompt Skill** writes one model-agnostic prompt spec and compiles it to whichever video model you can actually call. A spec records the decisions a prompt encodes — scope, locks, staging, end states — separately from the dialect that expresses them, so the same brief survives a change of model instead of being rewritten. Each model gets a measured profile covering reference syntax, limits, timing adherence, and default-bias behaviour; the Skill probes what it does not know, degrades the spec to what the model supports, and reports every degrade.
- **Which one to use:** use Seedance 2.5 Skill for Seedance-specific writing and execution. Use Universal Video Prompt Skill when one brief has to run across several models, when you are comparing models, or when the model you want is not available yet and the work has to proceed on another one.
- ### Install
- ```bash
- npx skills add AtlasCloudAI/awesome-seedance-2.5-prompts-skills --skill universal-video-prompt-skill
- ```
- This Skill can be installed on its own for standalone cross-model work. A complete Seedance 2.5 Skill setup installs all three Skills shown above.
- <details><summary>Skill files and references</summary>
- - [`SKILL.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/SKILL.md)
- - [`references/spec-format.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/references/spec-format.md)
- - [`references/verifiability.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/references/verifiability.md)
- - [`references/portability.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/references/portability.md)
- - [`references/film-type-dna.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/references/film-type-dna.md)
- - [`references/model-profile-schema.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/references/model-profile-schema.md)
- - [`references/execution.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/references/execution.md)
- - [`references/checklist.md`](https://github.com/AtlasCloudAI/awesome-seedance-2.5-prompts-skills/blob/main/skills/universal-video-prompt-skill/references/checklist.md)
- </details>
- <a id="how-to-use"></a>
- ## 🚀 How to use this repository
- 1. **Browse:** filter by category, open a real preview when available, and copy the prompt.
- 2. **Adapt:** ask Seedance 2.5 Skill to rewrite a prompt for your subject, duration, aspect ratio, references, and continuity needs.
- 3. **Generate:** let the Skill create and show intermediate storyboards when needed, review them automatically, and submit the final image or video request through the selected Atlas channel.
- <a id="model-and-execution-defaults"></a>
- ## ⚙️ Model and execution defaults
- The workflow Skill and the execution adapter are separate layers. The workflow decides what to make; the selected Atlas channel submits, polls, and retrieves the media.
- ### Model defaults
- **Storyboard stills:** Seedream 5.0 Pro. **Executable video default:** Seedance 2.0. **Seedance 2.5:** only when the selected provider exposes the model and its actual limits.
- - **Atlas Cloud Skill:** default direct execution route inside an Agent conversation. If it is missing, the workflow helps install it and directs the user to obtain an API key.
- - **Atlas MCP:** used when the user explicitly selects MCP and its generation tools are available.
- - **Atlas CLI / REST:** used for explicit terminal, script, CI, or batch workflows.
- All asynchronous jobs are polled every 2 seconds using the same prediction ID. A timeout or delayed output is not permission to submit a duplicate paid generation.
- **[→ Get an Atlas Cloud API key](https://www.atlascloud.ai/console/api-keys?utm_source=github&utm_campaign=awesome-seedance-2.5-prompts-skills)**
- <a id="curation-and-provenance"></a>
- ## 🔎 Curation and provenance
- Every prompt record keeps its category, source platform, author, source link, input references, and preview video when available. Prompt text remains unchanged during README generation.
- - The `official` and `community` labels describe where a prompt came from; they are not a guarantee that every prompt has been independently benchmarked across every provider or model version.
- - A preview demonstrates one observed output under its original setup. Results can change with model version, provider parameters, references, aspect ratio, seed, and moderation.
- <a id="statistics"></a>
- ## 📊 Statistics
- | Metric | Count |
- |--------|-------|
- | Total Prompts | **150** |
- | Categories | **33** |
- | Preview Videos | **50** |
- | Last Updated | **2026-11-18** |
- <a id="featured-prompts"></a>
- ## 🔥 Featured Prompts
- ### No. 1: Crystal Ball Match-Cut Beat Film
- - **Category:** `Match-Cut & Creative Effects`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group1-1.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group1-1.mp4" controls muted playsinline width="720"></video>
- #### Description
- A beat-synced match-cut film: one crystal ball etched with a glowing 'seedance' logo stays centered while eight scenes cut seamlessly behind it.
- #### Prompt
- ```text
- A fast-paced, cinematic Match-cut short film synced to a driving electronic beat. A flawless crystal ball stays fixed dead-center throughout, a glowing "seedance" logo etched inside it. The ball holds razor-sharp focus while, on every strong musical beat, the background match-cuts seamlessly: Scene 1: macro close-up, cinematic water splashing around the ball, refracting intricate light. Scene 2: a vintage morning cafe, the ball on a raw-wood table, rising coffee steam and blurred commuters beyond the window. Scene 3: golden-hour dusk, a skater youth tosses and catches the ball one-handed, street racing backward behind them in gorgeous backlit sunset. Scene 4: a frenzied music festival, hands raise the ball high, refracting dazzling stage lasers. Scene 5: a lively family party table, the ball resting center-frame, blurred figures toasting and reaching for food. Scene 6: a dim cinema, hands cupping the ball as the giant screen's faint glow drifts across its surface. Scene 7: the ball on a violently vibrating speaker diaphragm, match-cutting on the climax to a spinning DJ turntable center. Scene 8: an outdoor camping night, background becoming warm bonfire and swaying string-light bokeh. Finale: on the final downbeat the ball is hurled up out of frame; cut to pure black, a minimal white-on-black "seedance" appearing dead-center. Beat-synced match-cut editing, top-tier cinematic color grading, photoreal glass refraction, ray tracing, global illumination. Subject razor-sharp, background heavy motion blur.
- ```
- ### No. 2: Steampunk Clockwork — 30s One-Take
- - **Category:** `One-Take Cinematography`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group2-2.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group2-2.mp4" controls muted playsinline width="720"></video>
- #### Description
- A 30-second steampunk miniature 3D sequence with continuous orbiting, pass-through camera moves.
- #### Prompt
- ```text
- A high-end, deeply cinematic 30-second 3D motion-graphics sequence in refined steampunk and vintage-miniature style, with continuous fluid orbiting and pass-through camera moves. [0-10s] Macro close-up of an antique brass clock face that unfolds layer by layer into meshing rotating gear rings and volumetric fog. The camera pierces down through the gears; a mechanical ornithopter spirals up from a miniature canyon of stacked weathered old books. [10-20s] The camera glides forward tracking the ornithopter, seamlessly passing into a fast-spinning ornate brass zoetrope projecting galloping mechanical-horse light. The light leaps out and the scene becomes a brass floating cable car on glimmering copper rails through a forest of gears, bathed in cinematic golden-hour light. [20-30s] The camera pans elegantly down; below appears an exquisite clockwork wooden sailing ship cutting deep-blue glass-textured waves, which morph into a glowing giant moon with lantern-holding explorer silhouettes trekking a crystal-vein ridge under stars. The camera spirals smoothly back through ethereal clouds to the ticking brass clock face. Hyper-real mechanical textures, rich brass and gold tones, cinematic shallow depth of field, smooth seamless pass-through camerawork, epic fantastical adventure atmosphere.
- ```
- ### No. 3: Windows Through Worlds (5 refs)
- - **Category:** `Multi-Image Reference`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference1.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference2.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference3.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference4.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference5.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A brand concept short driven by 5 reference images, gliding through windows into different worlds.
- #### Prompt
- ```text
- A cinematic brand concept short. @image1 is the first frame; the picture trembles slightly, the camera pushes in to tree shadows rushing backward outside the window, their retreat accelerating, then abruptly cuts to @image2, speed easing as the camera glides slowly along a stream, birdsong and blossoms. The camera drops underwater — bubble sounds — as orange jellyfish drift gracefully past the lens @image3; the camera pulls back as small fish flit past and swim from the water into the window @image4, a girl looking around, watching them. The camera pulls back, defocuses, then refocuses sharp, switching to the music's rhythm: a Chinese-garden lattice window @image5 with light circling, church stained glass, an airplane porthole, a dome skylight, a bay window, louver blinds, a European dormer, a door peephole, a camera viewfinder, a bird's eye, a human eye close-up. It settles on the human eye; the eye closes, screen black, then suddenly opens — "seedance" appearing in the center of the eye on the accent beat.
- ```
- ### No. 4: One-Take Through Six Rooms (8 refs)
- - **Category:** `Multi-Image Reference`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference1.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference2.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference3.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference4.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference5.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference7.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference8.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference9.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- One continuous take following a figure through six mood-shifting rooms, driven by 8 reference images.
- #### Prompt
- ```text
- One continuous take. The camera smoothly follows a person in a black coat (reference @image1) moving left to right through six connected rooms of different tones and moods. Every room shares the same structure: white walls, light herringbone wood floor, French floor-to-ceiling windows, white sheer curtains (reference @image2) but the outside view and mood differ each time. The protagonist walks at a constant pace, passing through every open door. 0-5s room one, American-comic fight: the protagonist fights a character (@image3), who is defeated; 5-10s room two, warmth, felt-craft style, window view a sunflower field (@image4), warm-orange soft light, a painter painting sunflowers (@image5), the protagonist turning felt-style on entering; 10-15s room three, sorrow, black-and-white comic stop-motion, rain outside, cold-grey light, a person alone on the floor hugging their knees, a phone glowing with an unanswered call; on entering, the light blinks off then on, the room turns color, flowers bursting into bloom; 15-20s room four, joy, a room submerged in the sea (reference @image6), the protagonist turning transparent among coral and fish; 20-25s room five, surprise, window view a sky of fireworks (reference @image7), colorful flickering light, the protagonist swept up in a cheering crowd; 25-30s a blank white room, the protagonist snaps their fingers — snap SFX — frame goes black, "seedance" in the middle (reference @image8). Cinematic quality, high-fashion advertising style, lighting entirely determined by the window views for strong emotional contrast, no text in frame.
- ```
- ### No. 5: Video Edit — Remove All But the Lead
- - **Category:** `Video Editing`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group1-output.mov)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group1-output.mov" controls muted playsinline width="720"></video>
- - **Inputs:**
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group1-reference1.mp4" controls muted width="360"></video>
- #### Description
- Reference-video editing: erase everyone except the protagonist from a source clip.
- #### Prompt
- ```text
- Video editing: remove everyone in @video1 except the protagonist.
- ```
- ### No. 6: Multilingual Hip-Hop 'Hello' MV
- - **Category:** `Multi-Image Reference`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part3-group1-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part3-group1-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part3-group1-reference1.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A seaside hip-hop MV built from one reference image; the lead raps 'hello' in 8 languages, lip-synced, hard cuts on the beat.
- #### Prompt
- ```text
- Cinematic hip-hop / rap music video, photoreal quality, high-end tone, seaside setting. Build the frame from @image1: a band performs at a golden sand beach with crashing waves — a lead vocalist gripping a mic on a stand in the wet sand, one guitarist left, one right, a drummer at the back; a vast coastline behind, rolling waves, a warm golden-hour sun shimmering on the water, sea mist in the air. The lead in a red tracksuit raps to camera — lips and jaw precisely synced to every word, head punching to the beat. Bright, punchy, fast, confident rap. HARD CUT on the beat, each switch a double contrast (shot size and type change together). Lyrics (the lead sings 'hello' in each language in turn, precisely lip-synced): English "Hello", Chinese "你好", Japanese "こんにちは", Korean "안녕하세요", Portuguese "Olá", Thai "สวัสดี", Spanish "Hola", Arabic "مرحبا". 8 hard-cut shots (low-angle wide establishing; close-up rap to camera; macro guitar-string insert; 3/4 prowling orbit; lateral track at shore; drummer tilt-up; tight push on the lead; heroic full-band push-in), one language per shot. White balance 4000K, teal-and-amber grade, 35mm, shallow depth of field, film grain, sea mist, golden-hour flare. Premium feel, precise lip-sync, no subtitles, no text overlays, hard cuts only, total 20 seconds.
- ```
- ### No. 7: Capsule Coffee Machine Setup Tutorial (6 refs)
- - **Category:** `Reference-Guided Tutorial`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-output.mov)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-output.mov" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference1.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference2.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference3.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference4.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference5.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference6.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A 30-second setup-and-use tutorial for a capsule coffee machine, six reference-guided steps with voiceover — from installing the water tank to the first rinse.
- #### Prompt
- ```text
- A 30-second tutorial video on installing and using a capsule coffee machine. 0-2s: the opening title card reads "seedance capsule coffee machine setup tutorial". 2-5s, Step 1: install the water tank, reference @image1, medium shot from a slightly high angle, rear of the machine body, align the water tank with the slot on the back of the body and push it straight down until it clicks firmly into place, clearly showing how the tank's bottom clips align with the body's slot, the water level line visible through the tank's transparent section, voiceover "First, install the water tank. Align the tank with the slot on the back of the machine; a click means it is locked in place.". 5-9s, Step 2: install the drip tray, reference @image2, close-up front view, front bottom of the body, slide the drip tray horizontally into the guide rails at the bottom until fully seated, voiceover "Next, install the drip tray. Align the tray with the bottom rails.". 9-13s, Step 3: install the used-capsule collection box, reference @image3, close-up from a slightly low angle, the cavity beneath the drip tray, align the collection box with the recess and push it in flush with the drip tray, voiceover "Then insert the capsule collection box. Used capsules will drop down here automatically.". 13-18s, Step 4: first fill with water, reference @image4, close-up side view, the water tank at the top/back of the body, open the tank lid and pour in clean water up to the MAX water level line, then close it, emphasizing the water level line, voiceover "Open the tank lid, pour in clean water, being careful not to exceed the maximum water level line, then close the tank lid.". 18-25s, Step 5: power on, reference @image5, medium shot front view, front of the body, plug in the power cord and press the power button; the indicator light goes from blinking to steady (preheating complete), voiceover "Connect the power and press the power button. The indicator light starts blinking, which means it is preheating. When the light turns steady, the machine is ready.". 25-30s, Step 6: first rinse (without a capsule), reference @image6, medium shot moving to a close-up front-side view, without inserting a capsule press the brew button directly so hot water flows out and rinses the lines, emphasizing the "no capsule needed" note, voiceover "The last step, the first rinse. Note that this step does not require a capsule; just press the brew button. Once the rinse is done, your coffee machine is ready to use."
- ```
- ### No. 8: Fruity Cookie Commercial (image + 6 video refs)
- - **Category:** `Multimodal Reference (image+video)`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference1.png" width="180" referrerpolicy="no-referrer">
- </p>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference2.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference3.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference4.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference5.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference6.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference7.mp4" controls muted width="360"></video>
- #### Description
- A bright, colorful commercial for fruity cookies in four flavors — strawberry, apple, grape, orange — built from one image reference and six motion/camera video references, cut hard to a driving beat.
- #### Prompt
- ```text
- A bright, colorful commercial style with fruity cookies as the hero, in four flavors — strawberry, apple, grape, and orange; strawberry flavor reference @image1. The cookies and their matching fruits are arranged in a strongly ordered geometric array; the overall frame is clean, premium, and high-energy. It opens with the fruits orbiting rapidly around the central cookie to quickly establish visual focus — reference the composition of @video1 — cutting in on a strong musical beat. Then cookies of different flavors advance toward the lens along spiral paths, forming clear spatial depth — reference the motion and camera work of @video2 — switching colors and flavors on the beat with the background music. The array of cookies pans left and right with fast plane-to-plane cuts, strawberry, apple, grape, and orange flavors alternating as the frame jump-cuts quickly to the rhythm — reference the movement of @video3. The mid-section adds up-and-down panning; the neat cookie array rises and falls vertically like a machine — reference the movement of @video4 — highlighting the beauty of order and the richness of the product. In the climax a cookie is snapped in two and the moment enters slow motion as the fruity filling bursts open, crumbs scattering, the juicy sensation and grainy impact amplified — reference the explosion effect of @video5 — then quickly returns to the fast-paced edit. The ending brings in the English text "Fresh on Seedance, made for viral vision", entering word by word in quick succession with strong rhythmic text motion and a product freeze-frame — reference @video6 — the four cookie flavors lined up neatly with the fruits bouncing in sync for a final brand-forward close, the frame full of a young, energetic, delicious, shareable ad atmosphere.
- ```
- <a id="browse-by-category"></a>
- ## 🏷️ Browse by Category
- - [`Match-Cut & Creative Effects`](#category-1): **1**
- - [`One-Take Cinematography`](#category-2): **1**
- - [`Multi-Image Reference`](#category-3): **3**
- - [`Video Editing`](#category-4): **2**
- - [`Reference-Guided Tutorial`](#category-5): **1**
- - [`Multimodal Reference (image+video)`](#category-6): **1**
- - [`Image Reference`](#category-7): **1**
- - [`Image Reference (Cinematic)`](#category-8): **1**
- - [`E-commerce Livestreaming`](#category-9): **8**
- - [`Food & Dining`](#category-10): **8**
- - [`Beauty & Fashion`](#category-11): **6**
- - [`Digital Tech & Gadgets`](#category-12): **6**
- - [`Automotive`](#category-13): **4**
- - [`Real Estate & Home`](#category-14): **4**
- - [`Talking Head & Digital Humans`](#category-15): **6**
- - [`Short Dramas & Emotional Stories`](#category-16): **8**
- - [`Brand Commercials & Ads`](#category-17): **6**
- - [`Travel & Cultural Tourism`](#category-18): **6**
- - [`Pets & Animals`](#category-19): **5**
- - [`Kids, Education & Science`](#category-20): **5**
- - [`Gaming, CG & Fantasy`](#category-21): **5**
- - [`Animation & Cartoons`](#category-22): **4**
- - [`Sports & Fitness`](#category-23): **4**
- - [`Music, Dance & Music Videos`](#category-24): **4**
- - [`Festivals & Solar Terms`](#category-25): **5**
- - [`Nature, Astronomy & Documentaries`](#category-26): **4**
- - [`Anthropomorphic & Creative Memes`](#category-27): **2**
- - [`Film & Storytelling`](#category-28): **11**
- - [`Advertising & E-commerce`](#category-29): **8**
- - [`Knowledge & Explainer`](#category-30): **6**
- - [`Industrial & Manufacturing`](#category-31): **4**
- - [`Creative & Misc`](#category-32): **3**
- - [`Capability Showcase`](#category-33): **7**
- <a id="all-prompts"></a>
- ## 📋 All Prompts
- <a id="category-1"></a>
- ### Match-Cut & Creative Effects (1)
- ### No. 1: Crystal Ball Match-Cut Beat Film
- - **Category:** `Match-Cut & Creative Effects`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group1-1.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group1-1.mp4" controls muted playsinline width="720"></video>
- #### Description
- A beat-synced match-cut film: one crystal ball etched with a glowing 'seedance' logo stays centered while eight scenes cut seamlessly behind it.
- #### Prompt
- ```text
- A fast-paced, cinematic Match-cut short film synced to a driving electronic beat. A flawless crystal ball stays fixed dead-center throughout, a glowing "seedance" logo etched inside it. The ball holds razor-sharp focus while, on every strong musical beat, the background match-cuts seamlessly: Scene 1: macro close-up, cinematic water splashing around the ball, refracting intricate light. Scene 2: a vintage morning cafe, the ball on a raw-wood table, rising coffee steam and blurred commuters beyond the window. Scene 3: golden-hour dusk, a skater youth tosses and catches the ball one-handed, street racing backward behind them in gorgeous backlit sunset. Scene 4: a frenzied music festival, hands raise the ball high, refracting dazzling stage lasers. Scene 5: a lively family party table, the ball resting center-frame, blurred figures toasting and reaching for food. Scene 6: a dim cinema, hands cupping the ball as the giant screen's faint glow drifts across its surface. Scene 7: the ball on a violently vibrating speaker diaphragm, match-cutting on the climax to a spinning DJ turntable center. Scene 8: an outdoor camping night, background becoming warm bonfire and swaying string-light bokeh. Finale: on the final downbeat the ball is hurled up out of frame; cut to pure black, a minimal white-on-black "seedance" appearing dead-center. Beat-synced match-cut editing, top-tier cinematic color grading, photoreal glass refraction, ray tracing, global illumination. Subject razor-sharp, background heavy motion blur.
- ```
- <a id="category-2"></a>
- ### One-Take Cinematography (1)
- ### No. 1: Steampunk Clockwork — 30s One-Take
- - **Category:** `One-Take Cinematography`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group2-2.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group2-2.mp4" controls muted playsinline width="720"></video>
- #### Description
- A 30-second steampunk miniature 3D sequence with continuous orbiting, pass-through camera moves.
- #### Prompt
- ```text
- A high-end, deeply cinematic 30-second 3D motion-graphics sequence in refined steampunk and vintage-miniature style, with continuous fluid orbiting and pass-through camera moves. [0-10s] Macro close-up of an antique brass clock face that unfolds layer by layer into meshing rotating gear rings and volumetric fog. The camera pierces down through the gears; a mechanical ornithopter spirals up from a miniature canyon of stacked weathered old books. [10-20s] The camera glides forward tracking the ornithopter, seamlessly passing into a fast-spinning ornate brass zoetrope projecting galloping mechanical-horse light. The light leaps out and the scene becomes a brass floating cable car on glimmering copper rails through a forest of gears, bathed in cinematic golden-hour light. [20-30s] The camera pans elegantly down; below appears an exquisite clockwork wooden sailing ship cutting deep-blue glass-textured waves, which morph into a glowing giant moon with lantern-holding explorer silhouettes trekking a crystal-vein ridge under stars. The camera spirals smoothly back through ethereal clouds to the ticking brass clock face. Hyper-real mechanical textures, rich brass and gold tones, cinematic shallow depth of field, smooth seamless pass-through camerawork, epic fantastical adventure atmosphere.
- ```
- <a id="category-3"></a>
- ### Multi-Image Reference (3)
- ### No. 1: Windows Through Worlds (5 refs)
- - **Category:** `Multi-Image Reference`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference1.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference2.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference3.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference4.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/firstScreen-group3-reference5.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A brand concept short driven by 5 reference images, gliding through windows into different worlds.
- #### Prompt
- ```text
- A cinematic brand concept short. @image1 is the first frame; the picture trembles slightly, the camera pushes in to tree shadows rushing backward outside the window, their retreat accelerating, then abruptly cuts to @image2, speed easing as the camera glides slowly along a stream, birdsong and blossoms. The camera drops underwater — bubble sounds — as orange jellyfish drift gracefully past the lens @image3; the camera pulls back as small fish flit past and swim from the water into the window @image4, a girl looking around, watching them. The camera pulls back, defocuses, then refocuses sharp, switching to the music's rhythm: a Chinese-garden lattice window @image5 with light circling, church stained glass, an airplane porthole, a dome skylight, a bay window, louver blinds, a European dormer, a door peephole, a camera viewfinder, a bird's eye, a human eye close-up. It settles on the human eye; the eye closes, screen black, then suddenly opens — "seedance" appearing in the center of the eye on the accent beat.
- ```
- ### No. 2: One-Take Through Six Rooms (8 refs)
- - **Category:** `Multi-Image Reference`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference1.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference2.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference3.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference4.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference5.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference7.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference8.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group2-reference9.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- One continuous take following a figure through six mood-shifting rooms, driven by 8 reference images.
- #### Prompt
- ```text
- One continuous take. The camera smoothly follows a person in a black coat (reference @image1) moving left to right through six connected rooms of different tones and moods. Every room shares the same structure: white walls, light herringbone wood floor, French floor-to-ceiling windows, white sheer curtains (reference @image2) but the outside view and mood differ each time. The protagonist walks at a constant pace, passing through every open door. 0-5s room one, American-comic fight: the protagonist fights a character (@image3), who is defeated; 5-10s room two, warmth, felt-craft style, window view a sunflower field (@image4), warm-orange soft light, a painter painting sunflowers (@image5), the protagonist turning felt-style on entering; 10-15s room three, sorrow, black-and-white comic stop-motion, rain outside, cold-grey light, a person alone on the floor hugging their knees, a phone glowing with an unanswered call; on entering, the light blinks off then on, the room turns color, flowers bursting into bloom; 15-20s room four, joy, a room submerged in the sea (reference @image6), the protagonist turning transparent among coral and fish; 20-25s room five, surprise, window view a sky of fireworks (reference @image7), colorful flickering light, the protagonist swept up in a cheering crowd; 25-30s a blank white room, the protagonist snaps their fingers — snap SFX — frame goes black, "seedance" in the middle (reference @image8). Cinematic quality, high-fashion advertising style, lighting entirely determined by the window views for strong emotional contrast, no text in frame.
- ```
- ### No. 3: Multilingual Hip-Hop 'Hello' MV
- - **Category:** `Multi-Image Reference`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part3-group1-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part3-group1-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part3-group1-reference1.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A seaside hip-hop MV built from one reference image; the lead raps 'hello' in 8 languages, lip-synced, hard cuts on the beat.
- #### Prompt
- ```text
- Cinematic hip-hop / rap music video, photoreal quality, high-end tone, seaside setting. Build the frame from @image1: a band performs at a golden sand beach with crashing waves — a lead vocalist gripping a mic on a stand in the wet sand, one guitarist left, one right, a drummer at the back; a vast coastline behind, rolling waves, a warm golden-hour sun shimmering on the water, sea mist in the air. The lead in a red tracksuit raps to camera — lips and jaw precisely synced to every word, head punching to the beat. Bright, punchy, fast, confident rap. HARD CUT on the beat, each switch a double contrast (shot size and type change together). Lyrics (the lead sings 'hello' in each language in turn, precisely lip-synced): English "Hello", Chinese "你好", Japanese "こんにちは", Korean "안녕하세요", Portuguese "Olá", Thai "สวัสดี", Spanish "Hola", Arabic "مرحبا". 8 hard-cut shots (low-angle wide establishing; close-up rap to camera; macro guitar-string insert; 3/4 prowling orbit; lateral track at shore; drummer tilt-up; tight push on the lead; heroic full-band push-in), one language per shot. White balance 4000K, teal-and-amber grade, 35mm, shallow depth of field, film grain, sea mist, golden-hour flare. Premium feel, precise lip-sync, no subtitles, no text overlays, hard cuts only, total 20 seconds.
- ```
- <a id="category-4"></a>
- ### Video Editing (2)
- ### No. 1: Video Edit — Remove All But the Lead
- - **Category:** `Video Editing`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group1-output.mov)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group1-output.mov" controls muted playsinline width="720"></video>
- - **Inputs:**
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group1-reference1.mp4" controls muted width="360"></video>
- #### Description
- Reference-video editing: erase everyone except the protagonist from a source clip.
- #### Prompt
- ```text
- Video editing: remove everyone in @video1 except the protagonist.
- ```
- ### No. 2: Steam Train Bursts the Screen (video edit)
- - **Category:** `Video Editing`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group2-output.mov)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group2-output.mov" controls muted playsinline width="720"></video>
- - **Inputs:**
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part2-group2-reference1.mov" controls muted width="360"></video>
- #### Description
- A video edit that keeps the original @video1 old-cinema clip intact and adds only a steam locomotive bursting through the screen, tearing the fabric, with black-and-white snapping to full color at the moment of impact.
- #### Prompt
- ```text
- Hyper-real cinematic realist style with photorealistic live-action texture, emphasizing the real-world plausibility of the physical object bursting out (the steam train); no CGI gloss / game engine / stylized 3D. Preserve @video1's original composition, projection-hall lighting, handheld camera state and natural imperfections. Source video locked: keep @video1 fully intact as the base image (the old-fashioned projection hall, the rows of hat-wearing audience seen from behind, the projector's conical beam, the screen, the ambient light, the color tone and the handheld movement all unchanged); the only additions: the train bursting out of the screen, the changes in light and shadow it brings, the tearing of the screen, and the slight physical impact on the hall and the front-row audience. Color-shift rule (the core addition): at the start strictly keep @video1's original black-and-white old-film texture (grain / scratches / flicker / monochrome); at the exact moment the locomotive truly breaks through the screen and bursts into reality, the image abruptly shifts from black-and-white to full, real color, the color spreading across the whole frame like a shockwave centered on the point where the train breaks out (the train's black steel body, the red-copper boiler, the orange-yellow steam-lamp glow, the warm brown of the wooden carriage, the white-grey steam, and the warm brown of the projection hall all coming into color accordingly); the color shift is precisely synchronized with the screen-breaking burst, after which photorealistic color is maintained to the end. Screen-tearing locked: the screen is a real white projection cloth with genuine fabric tension; as the train bursts out it is forcibly torn open along the contour of the locomotive's front, with radial rips whose edges curl back and shredded cloth fibers flying. Subject: an old-fashioned steam locomotive front bursting forcibly out of the screen, a black steel body, a cylindrical boiler, a front cowcatcher, a chimney venting white-grey steam, a bright headlamp, the metal surfaces with real wear, oil grime, rivets and wet reflections, charging forward with immense weight and inertia. Camera: fully inherits @video1's original handheld movement, not smoothed, not re-timed, not re-composed; the bursting train stays correctly locked into the screen and the hall space at all times, maintaining correct parallax, occlusion and ground-contact shadow. Sound: no music, keep only real on-scene sound effects, inheriting @video1's projection-hall ambience (the projector's clatter, the audience's stirring) and adding the sound of the cloth tearing, steam gushing, the mechanical roar of steel wheels, and the impact air blast. Total duration exactly the same as @video1; add nothing other than the train itself and its physical interactions, the screen tearing, and the color shift at the moment of the break. No slow motion, no sense of magic, no stylized horror; the final result should look like an eerie yet real image of "a train smashing through the screen, black-and-white instantly turning to color" captured by chance by a real camera in an old projection hall.
- ```
- <a id="category-5"></a>
- ### Reference-Guided Tutorial (1)
- ### No. 1: Capsule Coffee Machine Setup Tutorial (6 refs)
- - **Category:** `Reference-Guided Tutorial`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-output.mov)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-output.mov" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference1.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference2.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference3.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference4.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference5.png" width="180" referrerpolicy="no-referrer">
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab1-group3-reference6.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A 30-second setup-and-use tutorial for a capsule coffee machine, six reference-guided steps with voiceover — from installing the water tank to the first rinse.
- #### Prompt
- ```text
- A 30-second tutorial video on installing and using a capsule coffee machine. 0-2s: the opening title card reads "seedance capsule coffee machine setup tutorial". 2-5s, Step 1: install the water tank, reference @image1, medium shot from a slightly high angle, rear of the machine body, align the water tank with the slot on the back of the body and push it straight down until it clicks firmly into place, clearly showing how the tank's bottom clips align with the body's slot, the water level line visible through the tank's transparent section, voiceover "First, install the water tank. Align the tank with the slot on the back of the machine; a click means it is locked in place.". 5-9s, Step 2: install the drip tray, reference @image2, close-up front view, front bottom of the body, slide the drip tray horizontally into the guide rails at the bottom until fully seated, voiceover "Next, install the drip tray. Align the tray with the bottom rails.". 9-13s, Step 3: install the used-capsule collection box, reference @image3, close-up from a slightly low angle, the cavity beneath the drip tray, align the collection box with the recess and push it in flush with the drip tray, voiceover "Then insert the capsule collection box. Used capsules will drop down here automatically.". 13-18s, Step 4: first fill with water, reference @image4, close-up side view, the water tank at the top/back of the body, open the tank lid and pour in clean water up to the MAX water level line, then close it, emphasizing the water level line, voiceover "Open the tank lid, pour in clean water, being careful not to exceed the maximum water level line, then close the tank lid.". 18-25s, Step 5: power on, reference @image5, medium shot front view, front of the body, plug in the power cord and press the power button; the indicator light goes from blinking to steady (preheating complete), voiceover "Connect the power and press the power button. The indicator light starts blinking, which means it is preheating. When the light turns steady, the machine is ready.". 25-30s, Step 6: first rinse (without a capsule), reference @image6, medium shot moving to a close-up front-side view, without inserting a capsule press the brew button directly so hot water flows out and rinses the lines, emphasizing the "no capsule needed" note, voiceover "The last step, the first rinse. Note that this step does not require a capsule; just press the brew button. Once the rinse is done, your coffee machine is ready to use."
- ```
- <a id="category-6"></a>
- ### Multimodal Reference (image+video) (1)
- ### No. 1: Fruity Cookie Commercial (image + 6 video refs)
- - **Category:** `Multimodal Reference (image+video)`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference1.png" width="180" referrerpolicy="no-referrer">
- </p>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference2.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference3.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference4.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference5.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference6.mp4" controls muted width="360"></video>
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab2-group2-reference7.mp4" controls muted width="360"></video>
- #### Description
- A bright, colorful commercial for fruity cookies in four flavors — strawberry, apple, grape, orange — built from one image reference and six motion/camera video references, cut hard to a driving beat.
- #### Prompt
- ```text
- A bright, colorful commercial style with fruity cookies as the hero, in four flavors — strawberry, apple, grape, and orange; strawberry flavor reference @image1. The cookies and their matching fruits are arranged in a strongly ordered geometric array; the overall frame is clean, premium, and high-energy. It opens with the fruits orbiting rapidly around the central cookie to quickly establish visual focus — reference the composition of @video1 — cutting in on a strong musical beat. Then cookies of different flavors advance toward the lens along spiral paths, forming clear spatial depth — reference the motion and camera work of @video2 — switching colors and flavors on the beat with the background music. The array of cookies pans left and right with fast plane-to-plane cuts, strawberry, apple, grape, and orange flavors alternating as the frame jump-cuts quickly to the rhythm — reference the movement of @video3. The mid-section adds up-and-down panning; the neat cookie array rises and falls vertically like a machine — reference the movement of @video4 — highlighting the beauty of order and the richness of the product. In the climax a cookie is snapped in two and the moment enters slow motion as the fruity filling bursts open, crumbs scattering, the juicy sensation and grainy impact amplified — reference the explosion effect of @video5 — then quickly returns to the fast-paced edit. The ending brings in the English text "Fresh on Seedance, made for viral vision", entering word by word in quick succession with strong rhythmic text motion and a product freeze-frame — reference @video6 — the four cookie flavors lined up neatly with the fruits bouncing in sync for a final brand-forward close, the frame full of a young, energetic, delicious, shareable ad atmosphere.
- ```
- <a id="category-7"></a>
- ### Image Reference (1)
- ### No. 1: Desert Horned Lizard & Grapefruit (3D ad)
- - **Category:** `Image Reference`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab3-group1-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab3-group1-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab3-group1-reference1.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A bright, glossy 3D animated commercial: a parched desert horned lizard finds a grapefruit, bites in, and a juice tsunami turns the desert into a summer sea. Built from one character reference.
- #### Prompt
- ```text
- A 3D animated commercial style, bright and translucent colors; the pulp and juice must feel intensely refreshing and impactful. The overall vibe is like a high-quality commercial animated short with a touch of exaggerated humor. The desert horned lizard character is cute, lively and expressive, reference @image1. The image texture references that soft natural light, delicate fuzz/skin texture, dreamy macro depth of field, and a realistic-yet-slightly-childlike feel from the reference image. 0-3s: a desert scorched by the blazing sun, the air shimmering with heat and the sand searing hot; a desert horned lizard lies on the scalding sand, tongue slightly out, eyes glazed, nearly dried out, swaying with every couple of steps as if about to "evaporate". 3-6s: it suddenly stops, its nose twitches, and looking down it spots a cool, plump, dew-covered grapefruit half-buried in the sand, glistening like a desert miracle; the lizard's eyes go wide in an instant. 6-8s: it pounces over, hugging the grapefruit tightly with both hands, pressing its whole face against the rind with a blissful "I'm finally alive" expression; the frame holds for 1 second, forming an exaggerated, funny memory beat. 8-11s: it looks up, opens its mouth and takes a big bite; the grapefruit rind splits open, the plump pulp shining with a translucent gloss, and the next moment the juice erupts like a tsunami. 11-16s: orange-pink, translucent, glistening grapefruit juice gushes out wildly, pouring down the dunes and rapidly flooding the entire desert; the dry yellow sand instantly turns into a cool, sparkling, fruit-scented summer sea, cacti, rocks and small dunes swallowed by the waves of juice, the lizard's expression turning from delight to terror. 16-20s: nearly drowned by the "grapefruit sea", it frantically clutches half a grapefruit like a life buoy and floats, poking its soaked head out looking dazed. 20-24s: cut to a white screen; the brand name and slogan appear dead-center: "Seedance Grapefruit — bite in for the pulp, what pours out is summer." The voiceover reads the whole line. 24-30s: cut back from white; the desert horned lizard is now lounging on the floating grapefruit, wearing tiny sunglasses and holding a cup with a straw, drifting leisurely on the "juice sea" on vacation, surrounded by floating orange pulp, little ice cubes, cool splashes and a clear blue sky; the mood shifts from "survival" to "vacation", and finally it leans back contentedly on the grapefruit as the camera pulls out and freezes on a refreshing, bright, playful summer frame.
- ```
- <a id="category-8"></a>
- ### Image Reference (Cinematic) (1)
- ### No. 1: Drunk Man's Rooftop Fall (black-comedy short)
- - **Category:** `Image Reference (Cinematic)`
- - **Source:** `official`
- - **Author:** Volcengine Ark
- - **Language:** `en`
- - **Video:** [View](https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab3-group2-output.mp4)
- <video src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab3-group2-output.mp4" controls muted playsinline width="720"></video>
- - **Inputs:**
- <p>
- <img src="https://static.atlascloud.ai/model/example/seedance-2.5/cn-part1-tab3-group2-reference1.png" width="180" referrerpolicy="no-referrer">
- </p>
- #### Description
- A ~30s black-comedy cinematic short: a chubby, tipsy man topples off a rooftop, grabs a lifeline, crashes through a window, and simply falls asleep. Character from one reference image; built on hard cuts between roar and total silence.
- #### Prompt
- ```text
- [Drunk Man's Rooftop Fright — falls off the edge — grabs a rope through the window — passes out on the spot | black-comedy short | about 30 seconds] One-line premise: a chubby, cute, drunk middle-aged man is eating a burger and drinking beer on the rooftop of a building; as he stands up and stretches he accidentally loses his footing and falls, in his panic grabs a lifesaving rope and immediately crashes through a window into an interior, is dazed and stunned for one second, then just lies down where he is and falls fast asleep. Reference image @image1 — a chubby, cute, drunk middle-aged man: round and stocky build with a big round beer belly, short in stature, movements loose, slow and wobbly from drinking; round face, double chin, drink-flushed cheeks and nose, eyes narrowed to slits with a bleary gaze, an untrimmed ring of short stubble, messy hair, a silly, cute, carefree expression; wearing a white tank-top undershirt stained with burger sauce hanging loosely over his round belly, old jeans, shuffling along in loose slippers, always clutching a half-eaten burger in his hand. There is only this one character in the whole film, a 100% match to the reference image. Overall look: a nighttime cyberpunk mega-city, a dense forest of glass-and-steel skyscrapers, neon signs (cyan / magenta / electric blue), holographic billboards, wet reflective ground, volumetric fog, distant flying-car light trails and drones. Real live-action texture, physical cinema lenses, no 3D rendering / game engine / anime feel. The camera mixes handheld with jib / drone; after the fall the camera plunges down with him as the building's exterior wall and the neon streak upward to convey speed, 24-35mm widescreen cinematic lens texture, 180-degree shutter motion blur. Sound design is the key to the black comedy: hard-cutting repeatedly between "a huge roar" and "total silence". Real gravity throughout. Storyboard (about 30 seconds): S1 (0-4s) opening chewing big close-up (with sound) the man's face fills the screen, chewing the burger in big bites, a satisfied drunken state; S2 (4-9s) pull to a wide shot, stretching -> losing footing and falling (with sound) he wobbles to his feet and stretches big, leans back, can't keep his balance, his foot slips and he tumbles backward over the edge of the rooftop, never letting go of the burger, with a panicked "Eh——!"; S3 (9-12s) cut to a startled facial close-up (a loud scream, huge roar) drunken eyes bulging round, double chin bunched up, mouth wide in a screaming "Ahh——!!", wind howling; S4 (12-16s) cut to a long shot at eye level (total silence) seen from the building across the way, a tiny white figure plummeting straight down from the high-rise, the neon city vast and indifferent, the total silence forming a black-comedy contrast with the huge roar a second earlier; S5 (16-21s) cut to a facial close-up (screaming) -> grabbing the rope, still screaming, hands flailing in the air, he suddenly grabs a lifesaving cable, the cable snapping taut and bending under his weight with a twang, the scream cutting off abruptly as he is jerked and swings out (a brief slow motion at the moment he grabs the rope); S6 (21-26s) cut to an interior angle (quiet) -> the glass shatters -> he crashes in, the interior first quiet with only the hum of the air conditioner, the next second the glass explosively bursts inward (a brief bullet time) as the drunk man, burger and all, smashes through the window, shards flying out radially, neon light streaming in, and he crashes heavily into the room onto the floor / sofa, broken glass clinking; S7 (26-30s) ending, a startled second -> falls asleep, the man sprawled spread-eagle on the floor, covered in broken glass, still clutching the burger, slowly opens his eyes, dazed, blank and startled for a full second, looking around as if thinking "where am I?", then the drunkenness takes over and, not caring at all, he smacks his lips, tucks the burger into his arms, rolls over, lies down right there and falls fast asleep, even snoring lightly, the camera pulls out and freezes, hard cut to black. Only on-set live sound, no music, no subtitles; the core is hard cuts repeatedly between "huge roar" and "silence".
- ```
- <a id="category-9"></a>
- ### E-commerce Livestreaming (8)
- ### No. 1: Garment On-Body Display
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the model from Image 1 (facial reference from Image 1 close-up, makeup and styling reference from Image 2 full-body shot) as the main subject, vertical 9:16, fashionable realistic style, generate a 30-second clothing display short video.
- #### Prompt
- ```text
- Using the model from Image 1 (facial reference from Image 1 close-up, makeup and styling reference from Image 2 full-body shot) as the main subject, vertical 9:16, fashionable realistic style, generate a 30-second clothing display short video.
- Shot 1: Medium shot, model @Image 1 stands sideways and suddenly turns, swinging the coat hem, pure color studio soft top light outlines the silhouette, camera slowly pushes in, (upbeat fashionable electronic music starts) <sound of fabric friction>.
- Shot 2: Close-up, camera smoothly pans across the shoulder seam and lapel stitching details, side backlight highlights the fabric sheen.
- Shot 3: Full shot, model @Image 1 naturally walks to showcase the overall silhouette and drape of the cut, camera smoothly follows.
- Shot 4: Medium shot, model @Image 1 raises hand to adjust collar and looks back, camera slowly orbits half a circle.
- Shot 5: Medium close-up, model @Image 1 freezes with a smile, camera slowly pulls out to end, (music concludes).
- Constraints: Model's face, makeup, and styling must remain consistent with Image 1 and Image 2 throughout; studio background and top light remain fixed.
- General Constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural slight movements, not be statues.
- ```
- ### No. 2: Footwear Details
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, high-end textured realistic style, main subject is a pair of genuine leather ankle boots, generate a 30-second footwear detail short video.
- #### Prompt
- ```text
- Vertical 9:16, high-end textured realistic style, main subject is a pair of genuine leather ankle boots, generate a 30-second footwear detail short video.
- Shot 1: Macro close-up, ankle boots resting on a matte stone platform, camera slowly pushes in to the toe stitching, cool-toned side light sweeps across the leather, (low atmospheric electronic music starts) <sound of light leather contact>.
- Shot 2: Close-up, camera orbits around the shoe body for one full circle showcasing the streamlined shape and heel curve, reflector fill light.
- Shot 3: Macro close-up, fixed camera position, fingertip flips the tongue to reveal the inner lining logo and stitching.
- Shot 4: Medium shot, booted foot steps down and lightly lifts the heel, camera looks down to capture the moment of contact, <sound of heel tapping ground>.
- Shot 5: Close-up, camera slowly pulls out to frame the full shoe, (music concludes).
- Constraints: Ankle boot style, leather texture, and color scheme must remain consistent throughout; stone platform and cool-toned lighting setup locked; prevent inconsistency between left and right shoe shapes.
- General Constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural slight movements, not be statues.
- ```
- ### No. 3: Bag Opening and Closing Demo
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, minimalist high-end realistic style, main subject is a commuter tote bag, generate a 30-second bag function demonstration short video.
- #### Prompt
- ```text
- Vertical 9:16, minimalist high-end realistic style, main subject is a commuter tote bag, generate a 30-second bag function demonstration short video.
- Shot 1: Medium shot, tote bag stands upright on a solid-color surface, camera slowly pushes in to the metal clasp, soft top light, (refreshing light music starts).
- Shot 2: Close-up, hands unzip to open the main compartment, camera looks down to showcase internal compartments and card slots, <sound of zipper sliding>.
- Shot 3: Close-up, hand places laptop and water bottle in sequence, camera smoothly pans to showcase capacity.
- Shot 4: Medium shot, single hand lifts the bag to showcase handle load-bearing and structured feel, camera low-angle shot.
- Shot 5: Medium shot, clasp is fastened and bag returns to position, camera slowly pulls out to end, (music concludes).
- Constraints: Tote bag color, hardware, and structure must remain consistent throughout; surface and top light locked; opening/closing actions should be连贯, not jumpy.
- General Constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural slight movements, not be statues.
- ```
- ### No. 4: Jewelry
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, luxurious and exquisite realistic style, main subject is a diamond necklace, generate a 30-second jewelry display short video.
- #### Prompt
- ```text
- Vertical 9:16, luxurious and exquisite realistic style, main subject is a diamond necklace, generate a 30-second jewelry display short video.
- Shot 1: Macro close-up, necklace coiled on a black velvet tray, camera slowly pushes in to the main diamond, spotlight creates fire, (elegant string music starts) <subtle metallic chime>.
- Shot 2: Close-up, camera orbits around the main diamond for half a circle capturing multi-angle refraction sparkles, background darkened.
- Shot 3: Medium close-up, necklace worn on model's neck, camera moves up from collarbone to the pendant, soft fill light on face.
- Shot 4: Macro close-up, fixed camera position, pendant sways slightly with breath, shimmering.
- Shot 5: Close-up, camera slowly pulls out focusing on the fire to frame, (string music concludes).
- Constraints: Necklace style, diamond cut, and setting must remain consistent throughout; black velvet background and spotlighting setup locked; neck skin tone in wearing segment should match studio segment.
- General Constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural slight movements, not be statues.
- ```
- ### No. 5: Home Appliance Display (Blender)
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, bright kitchen lifestyle realistic style, main subject is a high-speed blender, generate a 30-second home appliance demonstration short video.
- #### Prompt
- ```text
- Vertical 9:16, bright kitchen lifestyle realistic style, main subject is a high-speed blender, generate a 30-second home appliance demonstration short video.
- Shot 1: Medium shot, blender placed on a light-colored countertop, camera slowly pushes in to the control panel, morning light slants in, (lively lifestyle light music starts).
- Shot 2: Close-up, hand pours fruit and ice into the cup, camera looks down at the cup opening, <sound of ingredients falling>.
- Shot 3: Close-up, pressing the start button, blade assembly rotates at high speed to blend, camera fixed position close to the cup wall, <sound of blending roar>.
- Shot 4: Medium shot, pouring smooth smoothie into a glass, camera smoothly pans following the liquid flow.
- Shot 5: Medium close-up, finished smoothie and appliance body in the same frame, camera slowly pulls out to end, (music concludes).
- Constraints: Blender body color, panel, and cup must remain consistent throughout; kitchen countertop and morning light locked; smoothie texture should be连贯.
- General Constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural slight movements, not be statues.
- ```
- ### No. 6: Maternity and Baby Products
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, gentle and soothing lifestyle realistic style, main subject is a baby insulated feeding bottle, generate a 30-second maternity and baby product display short video.
- #### Prompt
- ```text
- Vertical 9:16, gentle and soothing lifestyle realistic style, main subject is a baby insulated feeding bottle, generate a 30-second maternity and baby product display short video.
- Shot 1: Medium close-up, feeding bottle standing on a nursery soft mat, camera slowly pushes in to the anti-colic nipple, warm soft light, (gentle piano music starts).
- Shot 2: Close-up, hand unscrews the cap to showcase wide-mouth design and measurement markings, camera smoothly pans, <gentle sound of cap unscrewing>.
- Shot 3: Close-up, after pouring warm water, hand touches bottle body to feel constant temperature, camera fixed position.
- Shot 4: Medium shot, mother holds bottle and gently shakes to test temperature before feeding, camera slowly moves up and down, <sound of water gently sloshing>.
- Shot 5: Medium close-up, feeding bottle and soft plush toy framed together, camera slowly pulls out to end, (music concludes).
- Constraints: Bottle color, body markings, and nipple style must remain consistent throughout; nursery warm light and soft mat background locked; maintain a soft, warm visual基调.
- General Constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural slight movements, not be statues.
- ```
- ### No. 7: Food Gift Box
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, festive and exquisite realistic style, featuring a Chinese pastry gift box. Generate a 30-second food gift box showcase short video.
- #### Prompt
- ```text
- Vertical 9:16, festive and exquisite realistic style, featuring a Chinese pastry gift box. Generate a 30-second food gift box showcase short video.
- Shot 1: Medium shot, a red and gold gift box rests quietly on a wooden table, the camera slowly pushes in to the gold-stamped cover, warm side lighting, (joyful Chinese-style music starts).
- Shot 2: Close-up, hands slowly lift the box lid to reveal the compartmentalized pastries, camera looks down for a full view, <sound of lid opening>.
- Shot 3: Macro close-up, the camera smoothly pans horizontally across the flaky crust and filling texture of each pastry compartment.
- Shot 4: Medium close-up, a hand picks up a piece of pastry to show its layers, the camera slowly orbits around it.
- Shot 5: Medium shot, the gift box is closed and tied with a ribbon, the camera slowly pulls back to conclude, (music concludes).
- Constraints: The gift box color scheme, gold-stamped pattern, and pastry types remain consistent throughout; the wooden table and warm lighting are locked; pastry placement is coherent from start to finish.
- General constraints: Keep no subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be statue-like.
- ```
- ### No. 8: Digital Accessories (Flash Sale Style, with Captions)
- - **Category:** `E-commerce Livestreaming`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, flash sale high-contrast realistic style, featuring a pair of wireless noise-cancelling headphones. Generate a 30-second promotional flash sale short video.
- #### Prompt
- ```text
- Vertical 9:16, flash sale high-contrast realistic style, featuring a pair of wireless noise-cancelling headphones. Generate a 30-second promotional flash sale short video.
- Shot 1: Close-up, headphones float and rotate against a neon gradient background, camera quickly pushes in, (strong rhythmic electronic drumbeat starts) <power-on prompt sound>, 【Storewide Direct Discount ¥300】.
- Shot 2: Macro close-up, camera cuts to the earcup hinge and charging contacts, hard light highlights sweep across.
- Shot 3: Medium shot, camera cuts to a person wearing headphones nodding to the beat, light and shadow flicker, 【Active Noise Cancellation 40-hour Battery Life】.
- Shot 4: Close-up, camera quickly pans horizontally past the charging case opening and closing, <magnetic snap sound>, 【Limited-time Flash Sale First 100 Customers】.
- Shot 5: Medium close-up, headphones and price tag share the frame, camera quickly pulls back and freezes, (drumbeat concludes) 【Buy Now】.
- Constraints: Headphone color and design remain consistent throughout; neon background color palette is locked; captions appear rhythmically synchronized with shot changes, maintaining high-contrast promotional feel.
- General constraints: Keep no subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be statue-like.
- ```
- <a id="category-10"></a>
- ### Food & Dining (8)
- ### No. 1: Hot Pot Served
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, steaming hot appetite-inducing realistic style, featuring a dual-flavor hot pot. Generate a 30-second hot pot serving short video.
- #### Prompt
- ```text
- Vertical 9:16, steaming hot appetite-inducing realistic style, featuring a dual-flavor hot pot. Generate a 30-second hot pot serving short video.
- Shot 1: Medium shot, the dual-flavor hot pot is placed on a wooden table, the red soup bubbling, camera slowly pushes in to the boiling oil bubbles, warm yellow restaurant lighting, (lively street market music starts) <soup base bubbling sound>.
- Shot 2: Close-up, a hand dips tripe into the red soup with the 'seven up, eight down' motion, camera looks down as steam rises, <blanching sound>.
- Shot 3: Macro close-up, camera smoothly pans horizontally across shrimp paste, freshly sliced beef, and a vegetable platter.
- Shot 4: Medium close-up, chopsticks lift a slice of meat coated in red oil, camera follows upward, <oil dripping sound>.
- Shot 5: Medium shot, several people raise their chopsticks together, creating a lively atmosphere, camera slowly pulls back to conclude, (music concludes).
- Constraints: The dual-flavor pot style, soup color, and dish arrangement remain consistent throughout; the restaurant's warm lighting and wooden table are locked; steam and boiling state persist throughout.
- General constraints: Keep no subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be statue-like.
- ```
- ### No. 2: Coffee Latte Art
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, tranquil and artistic realistic style, featuring a handcrafted latte with latte art. Generate a 30-second coffee latte art short video.
- #### Prompt
- ```text
- Vertical 9:16, tranquil and artistic realistic style, featuring a handcrafted latte with latte art. Generate a 30-second coffee latte art short video.
- Shot 1: Close-up, espresso drips into the cup, camera slowly pushes in to the crema texture, natural window light, (soothing jazz music starts) <espresso dripping sound>.
- Shot 2: Medium close-up, a hand holding a milk pitcher begins pouring milk foam, camera fixed at eye level looking at the liquid surface, <milk foam pouring sound>.
- Shot 3: Macro close-up, camera looks down as milk foam spreads on the coffee surface forming a leaf pattern.
- Shot 4: Close-up, a slight wrist flick finishes the pour, drawing the leaf veins, camera slowly orbits the rim of the cup.
- Shot 5: Medium shot, the finished latte is lifted and placed by the window, camera slowly pulls back to conclude, (music concludes).
- Constraints: The latte art pattern is coherent from pouring to completion; the window's natural light and tabletop are locked; milk foam texture is clear without breaks.
- General constraints: Keep no subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be statue-like.
- ```
- ### No. 3: Dessert Cut Open Revealing Lava (Molten Chocolate)
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, rich and tempting realistic style, featuring a molten chocolate lava cake. Generate a 30-second dessert lava flow short video.
- #### Prompt
- ```text
- Vertical 9:16, rich and tempting realistic style, featuring a molten chocolate lava cake. Generate a 30-second dessert lava flow short video.
- Shot 1: Medium close-up, the lava cake sits on a white porcelain plate dusted with powdered sugar, camera slowly pushes in, warm spotlight, (gentle romantic music starts).
- Shot 2: Close-up, a fork gently presses the cake surface creating a crack, camera fixed in a close position, <sound of crust lightly cracking>.
- Shot 3: Macro close-up, the moment a knife and fork cut open the cake, chocolate lava slowly oozes out, camera looks down for the climax moment, <thick sauce flowing sound>.
- Shot 4: Close-up, a fork lifts a piece with a string of lava, camera slowly moves up and down following the string.
- Shot 5: Medium shot, the lava cake and ice cream share the frame, camera slowly pulls back to conclude, (music concludes).
- Constraints: The cake's outer shell and lava color/texture remain consistent throughout; the white porcelain plate and warm spotlight are locked; the lava oozing out is the core money shot and must be clearly highlighted.
- General constraints: Keep no subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be statue-like.
- ```
- ### No. 4: Grilling on Charcoal
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Horizontal 16:9, rustic and smoky realistic style, featuring meat skewers grilled over charcoal fire. Generate a 30-second grilling short video.
- #### Prompt
- ```text
- Horizontal 16:9, rustic and smoky realistic style, featuring meat skewers grilled over charcoal fire. Generate a 30-second grilling short video.
- Shot 1: Close-up, meat skewers on a charcoal grill net, fat drips causing small flames, camera slowly pushes in, warm red charcoal fire light, (dynamic grilling atmosphere music starts) <sizzling charcoal sound>.
- Shot 2: Macro close-up, camera smoothly pans horizontally across the caramelized crispy crust of the skewers and sprinkled cumin and chili powder.
- Shot 3: Medium shot, a hand flips the skewers, creating rising smoke and flames, camera follows the hand movement, <sound of flipping and crackling>.
- Shot 4: Close-up, a basting brush sweeps sauce over the skewers as it seeps in, camera fixed in a close position.
- Shot 5: Medium close-up, the grilled skewers are arranged on a plate and lifted, camera slowly pulls back to conclude, (music concludes).
- Constraints: The type of meat skewers and degree of browning are coherent from start to finish; the charcoal fire and warm red light are locked; smoke and small flames persist to create a smoky atmosphere.
- General constraints: Keep no subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be statue-like.
- ```
- ### No. 5: Japanese Cuisine Plating
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, minimalist Zen realistic style, the main subject is a sashimi platter, generate a 30-second Japanese cuisine plating short video.
- #### Prompt
- ```text
- Vertical 9:16, minimalist Zen realistic style, the main subject is a sashimi platter, generate a 30-second Japanese cuisine plating short video.
- Shot 1: Medium shot, a black stone long plate is placed on a wooden tabletop, the camera slowly pushes in, cool top lighting, (clear and cold guzheng music starts).
- Shot 2: Close-up, hands use tweezers to precisely place tuna sashimi and shiso leaves, the camera shoots from a top-down angle for composition, <sound of ingredients being gently placed>.
- Shot 3: Macro close-up, the camera smoothly pans horizontally across the fat marbling of salmon and the sheen of ice grains.
- Shot 4: Medium close-up, drizzling soy sauce and garnishing with wasabi and petals, the camera slowly orbits around the finished dish.
- Shot 5: Full shot, the complete sashimi platter is centered and held still, the camera slowly pulls back to conclude, (music concludes).
- Constraints: The types of sashimi, slice thickness, and plating positions remain consistent throughout; the black stone plate and cool top lighting are locked; the composition maintains minimalist negative space.
- General constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural, slight movements, not be like statues.
- ```
- ### No. 6: Bubble Tea Making
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, fresh and vibrant realistic style, the main subject is a handcrafted pearl milk tea, generate a 30-second bubble tea making short video.
- #### Prompt
- ```text
- Vertical 9:16, fresh and vibrant realistic style, the main subject is a handcrafted pearl milk tea, generate a 30-second bubble tea making short video.
- Shot 1: Close-up, cooked brown sugar pearls are poured into the bottom of the cup, coating the sides, the camera slowly pushes in, bright bar lighting, (upbeat pop music starts) <sound of pearls falling into the cup>.
- Shot 2: Medium close-up, a hand-held milk pitcher pours freshly brewed tea and fresh milk, the camera is fixed at eye level showing the layering, <sound of pouring>.
- Shot 3: Macro close-up, the camera shoots from a top-down angle as ice cubes are dropped in, stirring the tea and milk, <sound of ice cubes clinking>.
- Shot 4: Close-up, a sealing machine presses the cup film and a straw is inserted, the camera follows the straw insertion action, <sound of piercing the film>.
- Shot 5: Medium shot, a hand holds up the finished milk tea and gently shakes it to show the pearls churning, the camera slowly pulls back to conclude, (music concludes).
- Constraints: The layering of the milk tea and the color of the pearls remain consistent throughout; the bright bar lighting and background are locked; the production steps are sequential and do not jump.
- General constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural, slight movements, not be like statues.
- ```
- ### No. 7: Hand-Pulled Noodles
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Horizontal 16:9, artisanal handcrafted realistic style, the main subject is handmade noodle pulling, generate a 30-second pulled noodles short video.
- #### Prompt
- ```text
- Horizontal 16:9, artisanal handcrafted realistic style, the main subject is handmade noodle pulling, generate a 30-second pulled noodles short video.
- Shot 1: Medium shot, the kneaded dough rests on the noodle board, hands press and fold it back, the camera slowly pushes in, warm kitchen lighting, (lively folk music starts) <sound of kneading dough>.
- Shot 2: Close-up, hands stretch and pull the dough, repeatedly folding and slapping it against the board, the camera follows the arc of the hands, <sound of slapping dough on the board>.
- Shot 3: Medium close-up, the noodles are pulled thinner and thinner into strands and dropped into a boiling pot, the camera shoots from a top-down angle showing the steam billowing, <sound of noodles entering boiling water>.
- Shot 4: Close-up, the noodles are scooped up and placed into a large bowl, then broth is poured over, the camera smoothly pans horizontally to show the soup.
- Shot 5: Medium shot, green onions are sprinkled and beef slices are arranged to complete the dish, the camera slowly pulls back to conclude, (music concludes).
- Constraints: The thickness of the noodles and the final toppings are consistent from start to finish; the warm kitchen lighting and noodle board are locked; the pulling action is continuous, reflecting the handmade feel.
- General constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural, slight movements, not be like statues.
- ```
- ### No. 8: Breakfast Scene (One-Take)
- - **Category:** `Food & Dining`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, warm and healing lifestyle realistic style, the main subject is a Chinese breakfast, generate a 30-second breakfast scene one-take short video.
- #### Prompt
- ```text
- Vertical 9:16, warm and healing lifestyle realistic style, the main subject is a Chinese breakfast, generate a 30-second breakfast scene one-take short video.
- Shot 1: One-take following shot, morning light streams through the curtains onto the dining table, the camera slowly pushes in on steaming soup dumplings, (gentle morning piano music starts) <sound of steamer lid being lifted>.
- Shot 2: The same long shot smoothly pans horizontally, gliding over golden fried eggs, youtiao (fried dough sticks), and soy milk, with steam and light spots flowing.
- Shot 3: The same long shot smoothly tilts up and down, a hand picks up a soup dumpling, dips it in vinegar, and lifts it, <sound of broth dripping>.
- Shot 4: The same long shot smoothly orbits a small half-circle, passing by an open book and a steaming coffee cup beside the table.
- Shot 5: The same long shot smoothly pulls back, framing a cozy full view of the breakfast, (music concludes).
- Constraints: Maintain a single, uninterrupted long take camera movement with no cuts throughout; the food arrangement and morning light tone remain consistent; the camera movement is smooth and connected, reflecting the one-take quality.
- General constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural, slight movements, not be like statues.
- ```
- <a id="category-11"></a>
- ### Beauty & Fashion (6)
- ### No. 1: Lipstick Swatch
- - **Category:** `Beauty & Fashion`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the lipstick in Image 1 (the bullet color and tube logo should match Image 1) as the main subject, vertical 9:16, high-saturation makeup aesthetic, generate a 30-second swatch short video.
- #### Prompt
- ```text
- Using the lipstick in Image 1 (the bullet color and tube logo should match Image 1) as the main subject, vertical 9:16, high-saturation makeup aesthetic, generate a 30-second swatch short video.
- Shot 1: Extreme close-up, the lipstick @Image 1 bullet slowly twists out, with velvet-like reflections flowing, the camera slowly pushes in, (crisp electronic music starts) <light sound of twisting the cap>.
- Shot 2: Close-up, the camera smoothly pans horizontally across the lips, the bullet glides across leaving a full, rich color.
- Shot 3: Medium close-up, low-angle shot of the model puckering her lips and smiling, the lip color shows under soft lighting, 【Today's Shade · True Red】.
- Shot 4: Close-up, the camera orbits the model's profile, the lip gloss highlight flows as she turns her head.
- Shot 5: Medium shot, the model holds the lipstick @Image 1 and poses towards the camera, the camera slowly pulls back, (music concludes).
- Constraints: The bullet color and tube logo must match Image 1 and not be changed; the lip color must match the bullet color.
- General constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural, slight movements, not be like statues.
- ```
- ### No. 2: Foundation Application
- - **Category:** `Beauty & Fashion`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the foundation bottle in Image 1 (the bottle and dropper should match Image 1) as the main subject, vertical 9:16, clean and translucent skincare-makeup finish, generate a 30-second application short video.
- #### Prompt
- ```text
- Using the foundation bottle in Image 1 (the bottle and dropper should match Image 1) as the main subject, vertical 9:16, clean and translucent skincare-makeup finish, generate a 30-second application short video.
- Shot 1: Extreme close-up, a drop of foundation @Image 1 hangs from the dropper and falls onto the back of the hand, the texture spreads, the camera slowly pushes in, <sound of liquid drop falling>.
- Shot 2: Close-up, the camera smoothly pans horizontally, a makeup sponge presses and blends on the cheek, the skin texture gradually becomes even.
- Shot 3: Medium close-up, following shot of the model turning her head, the cheek shows a flawless glow under natural light.
- Shot 4: Close-up, low-angle shot of the model with eyes closed during application, eyelashes cast soft shadows, (gentle piano music).
- Shot 5: Medium shot, the model holds the foundation bottle @Image 1 and smiles, posing, the camera slowly pulls back, 【Nude Flawless】.
- Constraints: The bottle and dropper must match Image 1; the makeup finish should look realistic without excessive airbrushing or plastic-like texture.
- General constraints: Keep no subtitles; Do not generate Logo; Do not generate watermarks; Use only one camera movement per shot; Characters should have natural, slight movements, not be like statues.
- ```
- ### No. 3: Perfume Ambiance
- - **Category:** `Beauty & Fashion`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the perfume bottle from Image 1 (bottle shape and engraving as per Image 1) as the main subject, horizontal 16:9, French retro ambiance lighting, generate a 30-second ambiance short film.
- #### Prompt
- ```text
- Using the perfume bottle from Image 1 (bottle shape and engraving as per Image 1) as the main subject, horizontal 16:9, French retro ambiance lighting, generate a 30-second ambiance short film.
- Shot 1: Close-up, perfume bottle @Image 1 stands on a stone windowsill, morning light passes through the bottle refracting light spots, camera slowly pushes in, (string ambiance music starts).
- Shot 2: Medium shot, camera slowly orbits around the bottle body, glass facets reflect flowing warm light.
- Shot 3: Close-up, low-angle shot of the spray moment, mist droplets suspend like golden dust in backlight, <spray light sound>.
- Shot 4: Medium close-up, follow the mist drifting towards a sheer curtain, curtain shadows sway gently with the wind.
- Shot 5: Medium shot, perfume bottle @Image 1 returns to the stone sill and freezes, camera slowly pulls out, (string music fades).
- Constraints: Bottle shape and body engraving must match Image 1; ambiance locked to warm-toned French retro, no cool colors.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural slight movement, not statues.
- ```
- ### No. 4: Skincare Essence Dropping
- - **Category:** `Beauty & Fashion`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the essence bottle from Image 1 (bottle body and dropper scale as per Image 1) as the main subject, vertical 9:16, crystal clear high-moisture texture, generate a 30-second texture short film.
- #### Prompt
- ```text
- Using the essence bottle from Image 1 (bottle body and dropper scale as per Image 1) as the main subject, vertical 9:16, crystal clear high-moisture texture, generate a 30-second texture short film.
- Shot 1: Extreme close-up, dropper suspended in mid-air, a drop of essence @Image 1 is about to fall, camera fixed position, <subtle bubble sound>.
- Shot 2: Extreme close-up, essence drop falls and hits the water surface, slow ripples spread, camera slowly pushes in, (water droplet musical accent).
- Shot 3: Close-up, camera smoothly pans horizontally, essence forms a crystal clear thread stretching from fingertips.
- Shot 4: Medium close-up, overhead shot of essence being applied to the back of the hand, gloss spreads like a mirror.
- Shot 5: Medium shot, essence bottle @Image 1 stands frozen against a water-light background, camera slowly rises, 【24-hour hydration lock】.
- Constraints: Bottle body and dropper scale must match Image 1; texture must be translucent and stringy, not cloudy.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural slight movement, not statues.
- ```
- ### No. 5: Fashion Street Snap
- - **Category:** `Beauty & Fashion`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the model's styling from Image 1 (clothing and accessories as per Image 1) as the main subject, vertical 9:16, film grain urban street feel, generate a 30-second street snap short film.
- #### Prompt
- ```text
- Using the model's styling from Image 1 (clothing and accessories as per Image 1) as the main subject, vertical 9:16, film grain urban street feel, generate a 30-second street snap short film.
- Shot 1: Medium shot, model @Image 1 pushes open a glass door and steps onto the street, wind lifts the hem of the clothes, camera follows, (retro disco rhythm starts).
- Shot 2: Full shot, camera smoothly pans horizontally following the model walking past neon shop windows, reflections flow.
- Shot 3: Medium close-up, low-angle shot of the model looking back and taking off sunglasses, backlight outlines golden edges of hair.
- Shot 4: Close-up, camera slowly pushes in on accessory details, metal reflections flicker.
- Shot 5: Full shot, model @Image 1 stands still in the center of the street, arms outstretched and freezes, camera slowly pulls out and rises, 【CITY WALK】(music concludes).
- Constraints: Clothing and accessories must match Image 1, do not alter styles; film grain and urban neon color tone locked throughout.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural slight movement, not statues.
- ```
- ### No. 6: Hairstyle Showcase
- - **Category:** `Beauty & Fashion`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the model's hairstyle from Image 1 (hair color and style as per Image 1) as the main subject, vertical 9:16, salon-advertisement-level soft light texture, generate a 30-second hairstyle short film.
- #### Prompt
- ```text
- Using the model's hairstyle from Image 1 (hair color and style as per Image 1) as the main subject, vertical 9:16, salon-advertisement-level soft light texture, generate a 30-second hairstyle short film.
- Shot 1: Close-up, model @Image 1 lowers head in anticipation, hair strands fall covering the face, camera fixed position, (rhythmic music intro).
- Shot 2: Medium close-up, model suddenly tosses head, hair strands arc up in slow motion, camera slowly pulls out, <hair toss wind sound>.
- Shot 3: Close-up, camera orbits around the head, light and shadow sweep across the smooth highlights of the hair.
- Shot 4: Medium shot, follow the model's fingers running through the hair, hair layers fall distinctly.
- Shot 5: Medium close-up, model @Image 1 turns and looks back frozen, hairstyle full and set, camera slowly pushes in, 【Mirror-like shine】.
- Constraints: Hair color and style silhouette must match Image 1; hair toss must be natural and smooth, no clipping or breakage.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural slight movement, not statues.
- ```
- <a id="category-12"></a>
- ### Digital Tech & Gadgets (6)
- ### No. 1: Smartphone Product Rotation
- - **Category:** `Digital Tech & Gadgets`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the smartphone from Image 1 (body and logo as per Image 1) as the main subject, horizontal 16:9, tech high-end cool tone, generate a 30-second product showcase film.
- #### Prompt
- ```text
- Using the smartphone from Image 1 (body and logo as per Image 1) as the main subject, horizontal 16:9, tech high-end cool tone, generate a 30-second product showcase film.
- Shot 1: Extreme close-up, smartphone @Image 1 on a dark surface lights up, cool light outlines its edges, camera slowly pushes in, (low-frequency electronic sound starts).
- Shot 2: Medium shot, camera smoothly orbits to showcase the smartphone @Image 1 backplate metal texture and camera module.
- Shot 3: Close-up, camera slowly pans horizontally across side button and port details.
- Shot 4: Medium close-up, low-angle shot of the smartphone @Image 1 slowly standing up and levitating, cool light spreads up from the bottom.
- Shot 5: Medium shot, smartphone @Image 1 levitates and rotates frozen, camera slowly pulls out, (music concludes).
- Constraints: Body and logo must match Image 1, do not alter design; do not fabricate interface content; cool-toned industrial lighting locked throughout.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural slight movement, not statues.
- ```
- ### No. 2: Earbuds Ejecting from Charging Case
- - **Category:** `Digital Tech & Gadgets`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the earbud charging case from Image 1 (shape and logo as per Image 1) as the main subject, horizontal 16:9, minimalist pure white studio shoot texture, generate a 30-second unboxing short film.
- #### Prompt
- ```text
- Using the earbud charging case from Image 1 (shape and logo as per Image 1) as the main subject, horizontal 16:9, minimalist pure white studio shoot texture, generate a 30-second unboxing short film.
- Shot 1: Extreme close-up, charging case @Image 1 rests on a pure white surface, fingertip approaches the lid, camera fixed position, (crisp prompt sound).
- Shot 2: Close-up, lid slowly opens, indicator light turns on, camera slowly pushes in, <opening/closing latch sound>.
- Shot 3: Medium close-up, earbuds slowly rise and levitate from the case, camera slowly lifts and follows.
- Shot 4: Close-up, camera orbits a single earbud, showcasing the in-ear angle and microphone openings.
- Shot 5: Medium shot, earbuds and charging case @Image 1 placed together frozen, camera slowly pulls out, 【Seamless portability】.
- Constraints: Shape and logo must match Image 1, do not alter structure; pure white seamless background locked, no added clutter.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural slight movement, not statues.
- ```
- ### No. 3: Notebook Opening and Closing
- - **Category:** `Digital Tech & Gadgets`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the laptop from Image 1 (case and logo as per Image 1) as the main subject, landscape 16:9, premium metallic cold-light texture, generate a 30-second showcase video.
- #### Prompt
- ```text
- Using the laptop from Image 1 (case and logo as per Image 1) as the main subject, landscape 16:9, premium metallic cold-light texture, generate a 30-second showcase video.
- Shot 1: Extreme close-up, the notebook @Image 1 lies on its side with the lid closed, cold light sweeps over the aluminum alloy hinge, the camera slowly pans horizontally, (low-frequency ambient sound starts).
- Shot 2: Medium close-up, the screen hinge slowly opens and closes until upright, the screen gradually brightens, the camera slowly pushes in, <hinge damping sound>.
- Shot 3: Close-up, overhead shot of the keyboard backlight lighting up key by key, a wave of light flows through.
- Shot 4: Medium shot, the camera smoothly orbits to showcase the ultra-thin side profile and ports of the body.
- Shot 5: Medium shot, the notebook @Image 1 is suspended in a half-open pose, frozen in place, the camera slowly pulls out, (music concludes).
- Constraints: Case thickness and logo as per Image 1; do not fabricate specific screen UI content; metallic cold light locked throughout.
- General Constraints: Keep without subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural, subtle movement, not like statues.
- ```
- ### No. 4: Smart Watch Interface
- - **Category:** `Digital Tech & Gadgets`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the smart watch from Image 1 (case and strap as per Image 1) as the main subject, portrait 9:16, bright and lively tech feel, generate a 30-second feature video.
- #### Prompt
- ```text
- Using the smart watch from Image 1 (case and strap as per Image 1) as the main subject, portrait 9:16, bright and lively tech feel, generate a 30-second feature video.
- Shot 1: Extreme close-up, the watch @Image 1 raises wrist to wake screen, animation expands, the camera slowly pushes in, (bright electronic music starts).
- Shot 2: Close-up, the camera slowly pans horizontally, the crown rotates to switch watch face cards.
- Shot 3: Medium close-up, follow shot of wrist swinging, activity data rings jump with the movement.
- Shot 4: Close-up, orbiting shot showcasing the texture of the case side and strap clasp.
- Shot 5: Medium shot, the watch @Image 1 is worn on the wrist, frozen in place as it lifts, the camera slowly pulls out, 【All-day protection】.
- Constraints: Case and strap as per Image 1; interface animations should be simple, do not fabricate specific numerical details.
- General Constraints: Keep without subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural, subtle movement, not like statues.
- ```
- ### No. 5: Drone Flight
- - **Category:** `Digital Tech & Gadgets`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the drone from Image 1 (body and propellers as per Image 1) as the main subject, landscape 16:9, expansive aerial documentary feel, generate a 30-second flight video.
- #### Prompt
- ```text
- Using the drone from Image 1 (body and propellers as per Image 1) as the main subject, landscape 16:9, expansive aerial documentary feel, generate a 30-second flight video.
- Shot 1: Medium close-up, the drone @Image 1 slowly lifts off from grass, propellers blurred at high speed, the camera slowly lifts and follows, <propeller hum> (ambient music starts).
- Shot 2: Close-up, the camera orbits the hovering body, the gimbal slightly adjusts the lens direction.
- Shot 3: Wide shot, follow shot of the drone skimming over a lake surface, its reflection gliding along.
- Shot 4: Extreme long shot, overhead shot of the drone flying over a mountain ridge, the land spreading out below.
- Shot 5: Medium shot, the drone @Image 1 returns smoothly and hovers, frozen in place, the camera slowly pulls out, (music concludes).
- Constraints: Body and propeller structure as per Image 1; flight posture must be stable, conforming to real aerodynamics, no drifting or shaking.
- General Constraints: Keep without subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural, subtle movement, not like statues.
- ```
- ### No. 6: Chip Macro (Tech Documentary)
- - **Category:** `Digital Tech & Gadgets`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the chip from Image 1 (package and silkscreen as per Image 1) as the main subject, landscape 16:9, cold, tech documentary macro texture, generate a 30-second macro video.
- #### Prompt
- ```text
- Using the chip from Image 1 (package and silkscreen as per Image 1) as the main subject, landscape 16:9, cold, tech documentary macro texture, generate a 30-second macro video.
- Shot 1: Extreme macro close-up, the chip @Image 1 rests on a circuit board, gold pins gleam with cold light, the camera slowly pushes in, (low-frequency industrial sound starts).
- Shot 2: Macro close-up, the camera smoothly pans across the package surface silkscreen and cutting edges.
- Shot 3: Macro close-up, overhead shot of solder joints and traces, shallow depth of field with only a line in focus.
- Shot 4: Medium close-up, the camera slowly pulls out to reveal the entire circuit board array, indicator lights flicker.
- Shot 5: Close-up, the chip @Image 1 is frozen in cold light, the camera slowly orbits to conclude, (music fades out).
- Constraints: Package model and silkscreen as per Image 1, do not alter; macro depth of field must be realistic, do not fabricate non-existent circuit structures.
- General Constraints: Keep without subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural, subtle movement, not like statues.
- ```
- <a id="category-13"></a>
- ### Automotive (4)
- ### No. 1: Night Scene Follow Shot
- - **Category:** `Automotive`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the car from Image 1 (model and paint as per Image 1) as the main subject, landscape 16:9, cyberpunk night scene cinematic texture, generate a 30-second follow-shot video.
- #### Prompt
- ```text
- Using the car from Image 1 (model and paint as per Image 1) as the main subject, landscape 16:9, cyberpunk night scene cinematic texture, generate a 30-second follow-shot video.
- Shot 1: Medium close-up, the car @Image 1 stops at a neon-lit street intersection, headlights illuminate, rain-slicked surface reflects light colors, the camera slowly pushes in, (deep electronic beat starts) <engine gentle roar>.
- Shot 2: Medium shot, side follow shot of the vehicle driving down a long neon street, light bands flow over the car paint.
- Shot 3: Close-up, the camera smoothly pans across the front fascia headlights and grille cold light.
- Shot 4: Wide shot, low-angle shot of the vehicle passing through a tunnel, overhead strip lights quickly sweep by.
- Shot 5: Medium shot, the car @Image 1 exits the tunnel, slows down and freezes, the camera slowly pulls out and rises, (rhythm concludes).
- Constraints: Car model and paint color as per Image 1, do not alter; night scene neon and wet surface reflective atmosphere locked throughout.
- General Constraints: Keep without subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural, subtle movement, not like statues.
- ```
- ### No. 2: Detail Close-up Quick Cuts
- - **Category:** `Automotive`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the car from Image 1 (model and paint as per Image 1) as the main subject, landscape 16:9, bold high-contrast industrial texture, generate a 30-second detail video.
- #### Prompt
- ```text
- Using the car from Image 1 (model and paint as per Image 1) as the main subject, landscape 16:9, bold high-contrast industrial texture, generate a 30-second detail video.
- Shot 1: Extreme close-up, the car @Image 1 headlights illuminate the light strip, cold light outlines the edges, the camera slowly pushes in, (urgent drumbeat starts).
- Shot 2: Close-up, cut to the wheel rim, slowly rotating, reflecting light spots.
- Shot 3: Close-up, cut to the emblem and grille, metallic highlights sweep across.
- Shot 4: Close-up, cut to the exhaust and spoiler lines, the camera smoothly pans.
- Shot 5: Medium shot, the car @Image 1 entire vehicle frozen in a beam of light, the camera slowly pulls out, (drumbeat concludes) 【Every detail is sharp】.
- Constraints: Car model and paint as per Image 1; quick cuts, each shot focuses on a single component without repetition; high-contrast hard light locked throughout.
- General Constraints: Keep without subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural, subtle movement, not like statues.
- ```
- ### No. 3: Off-Road Scene
- - **Category:** `Automotive`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the off-road vehicle from Image 1 (model and paint color as per Image 1) as the main subject, landscape 16:9, rough wilderness documentary feel, generate a 30-second off-road short film.
- #### Prompt
- ```text
- Using the off-road vehicle from Image 1 (model and paint color as per Image 1) as the main subject, landscape 16:9, rough wilderness documentary feel, generate a 30-second off-road short film.
- Shot 1: Medium shot, the off-road vehicle @Image 1 charges up a sand dune kicking up a dust wave, camera follows, <sound of tires grinding sand> (rousing drum music begins).
- Shot 2: Close-up, low-angle shot of wheels crushing over gravel, mud and sand splattering onto the lens.
- Shot 3: Wide shot, smooth lateral tracking shot following the vehicle fording a shallow stream, water splashing.
- Shot 4: Extreme wide shot, overhead shot of the off-road vehicle traversing tracks on the wasteland, a long trail of dust.
- Shot 5: Medium shot, the off-road vehicle @Image 1 reaches the ridge top and stops for a final frame, camera slowly rises and pulls back, (drum music concludes).
- Constraints: Vehicle model and paint color must match Image 1; off-road dynamics must conform to realistic terrain physics, no floating, slipping, or distortion.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural subtle movement, not stone statues.
- ```
- ### No. 4: Interior Showcase
- - **Category:** `Automotive`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Using the car interior from Image 1 (materials and color scheme as per Image 1) as the main subject, landscape 16:9, warm and premium cabin lighting feel, generate a 30-second interior short film.
- #### Prompt
- ```text
- Using the car interior from Image 1 (materials and color scheme as per Image 1) as the main subject, landscape 16:9, warm and premium cabin lighting feel, generate a 30-second interior short film.
- Shot 1: Extreme close-up, the leather seat stitching @Image 1 undulates under warm light, camera slowly pans, (soothing ambient music begins).
- Shot 2: Medium close-up, camera slowly pushes in towards the central control screen, ambient lighting gradually illuminates.
- Shot 3: Close-up, orbiting around the steering wheel and paddle shifters, metallic trim reflections flow.
- Shot 4: Medium shot, overhead shot of the full cabin, skylight light pours down from above.
- Shot 5: Medium close-up, camera slowly rises ending on a final frame of the cabin, warm light suffuses, (music fades) 【Serene Cabin】.
- Constraints: Materials and color scheme must match Image 1, no changes; do not fabricate specific UI content for the central control screen; warm cabin lighting tone locked throughout.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural subtle movement, not stone statues.
- ```
- <a id="category-14"></a>
- ### Real Estate & Home (4)
- ### No. 1: Living Room Space Walkthrough
- - **Category:** `Real Estate & Home`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, real estate quality promotional film, bright, airy, warm white tone, generate a 30-second living room space walkthrough short film.
- #### Prompt
- ```text
- Vertical 9:16, real estate quality promotional film, bright, airy, warm white tone, generate a 30-second living room space walkthrough short film.
- Shot 1: Fixed camera position, morning light slants into the empty living room from floor-to-ceiling windows, dust motes float in the light beams, (gentle piano begins).
- Shot 2: Smooth pan, camera glides close to the ground over wooden flooring and a beige fabric sofa, pillow textures clear, <soft footsteps>.
- Shot 3: Slow push-in, focus on a steaming cup of coffee and an open book on the coffee table, warm light falls, (strings layer in).
- Shot 4: Crane up, camera position slowly lifts from the ground, revealing the open layout connecting the living room and open-plan kitchen/dining area, 【Floor Area 98㎡ North-South Through Ventilation】.
- Shot 5: Slow pull-back, retreats to the entrance hall looking back at the depth of the entire living room, warm white final frame, (music concludes).
- Constraints: No characters appear throughout; furniture and decor style unified as modern minimalist wood style; lighting maintains consistent morning warm white.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural subtle movement, not stone statues.
- ```
- ### No. 2: Kitchen Details
- - **Category:** `Real Estate & Home`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, high-end appliance texture advertisement, cool metallic gray paired with warm wood tones, generate a 30-second kitchen details short film.
- #### Prompt
- ```text
- Vertical 9:16, high-end appliance texture advertisement, cool metallic gray paired with warm wood tones, generate a 30-second kitchen details short film.
- Shot 1: Close-up, water pours from the faucet into the stainless steel sink, water droplets splash in slow motion, <sound of flowing water>.
- Shot 2: Smooth pan, camera slides along the quartz countertop past knives, a cutting board, and fresh vegetables, cold light highlights, (lively rhythm begins).
- Shot 3: Slow push-in, a drawer automatically slides out showcasing compartmentalized storage and matte handles, warm wood tones, <sound of smooth drawer gliding>.
- Shot 4: Overhead shot, gas stove blue flame ignites, oil sizzles in the pan, steam rises, <click of ignition><sizzle of oil>.
- Shot 5: Low-angle shot, camera position looks up from the stove towards the range hood and recessed light strips, metallic reflections, 【Integrated Built-in Kitchen Appliances】, (music concludes).
- Constraints: No characters appear throughout; countertops remain tidy and clutter-free; metal and wood material textures consistent throughout.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural subtle movement, not stone statues.
- ```
- ### No. 3: Floor Plan Overhead (Schematic)
- - **Category:** `Real Estate & Home`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Landscape 16:9, architectural visualization schematic style, clean white model with soft skylight, generate a 30-second floor plan overhead schematic short film.
- #### Prompt
- ```text
- Landscape 16:9, architectural visualization schematic style, clean white model with soft skylight, generate a 30-second floor plan overhead schematic short film.
- Shot 1: Overhead shot, directly vertical top-down view of the complete floor plan white model, partition lines clear, (minimal electronic sound begins).
- Shot 2: Slow push-in, camera descends through the "vanished ceiling" into the living room schematic area, furniture presented semi-transparently, <space transition sound effect>.
- Shot 3: Orbit, rotates once around the master bedroom schematic area, showcasing bay window and walk-in closet circulation, soft light gradually brightens, (rhythm intensifies).
- Shot 4: Smooth pan, slides along the corridor schematic circulation towards the secondary bedroom and bathroom, annotations appear, 【Three Bedrooms Two Living Rooms Dynamic and Static Zoning】.
- Shot 5: Crane up, camera position rises back to the global overhead view with area data superimposed, skylight final frame, (music concludes).
- Constraints: Entirely schematic white model, not real scene, must include the annotation "Schematic for reference only"; floor plan proportion and structure consistent throughout; no real people or furniture.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural subtle movement, not stone statues.
- ```
- ### No. 4: Bedroom Ambiance
- - **Category:** `Real Estate & Home`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Vertical 9:16, slow living ambiance advertisement, warm orange night light soft tone, generate a 30-second bedroom ambiance short film.
- #### Prompt
- ```text
- Vertical 9:16, slow living ambiance advertisement, warm orange night light soft tone, generate a 30-second bedroom ambiance short film.
- Shot 1: Fixed camera position, bedside warm light turns on instantly illuminating plain bedding and half-drawn curtains, (ambient music begins).
- Shot 2: Slow push-in, camera moves close to the pillow, gliding over soft quilt folds and an open book, light and shadow gradually warm, <sound of light rain outside>.
- Shot 3: Smooth pan, slides along the foot of the bed past plants and a diffuser on the bay window, warm orange light halo flows, (gentle piano notes).
- Shot 4: Crane up, camera position slowly lifts for an overhead view of the entire bed and pendant light soft glow, creating a sense of sleeping envelopment, 【Bedroom with Excellent Natural Light and Ventilation】.
- Shot 5: Slow pull-back, retreats to the doorway looking back at the bedroom panorama under the warm light, soft tone final frame, (music concludes).
- Constraints: No characters appear throughout; lighting maintains consistent warm orange night tone; bedding and decor maintain unified elegant style.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters need natural subtle movement, not stone statues.
- ```
- <a id="category-15"></a>
- ### Talking Head & Digital Humans (6)
- ### No. 1: Knowledge Monologue
- - **Category:** `Talking Head & Digital Humans`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The host is bound to @Image 1 (facial makeup and styling should correspond to the image), vertical screen 9:16, clean studio style for a knowledge blogger with bright soft light, generate a 30-second knowledge monologue short video.
- #### Prompt
- ```text
- The host is bound to @Image 1 (facial makeup and styling should correspond to the image), vertical screen 9:16, clean studio style for a knowledge blogger with bright soft light, generate a 30-second knowledge monologue short video.
- Shot 1: Medium close-up, host @Image 1 leans forward looking directly at the camera to pose a question, shallow depth of field bookshelf background, fixed camera position, {Did you know 90% of people use it wrong?}.
- Shot 2: Slow push-in, host @Image 1 speaks while raising a finger for emphasis, determined gaze, warm light brightening, (upbeat background music starts).
- Shot 3: Cut to, key point graphics appear on the right side of the screen, host @Image 1 turns sideways and points, smooth pan, 【Key Point One Two Three】.
- Shot 4: Medium close-up, host @Image 1 turns back to face front, smiling for the summary, tone slows down, fixed camera position, {Remember these three points and that's enough}.
- Shot 5: Slow pull-out, host @Image 1 nods to conclude and waves, soft light freeze frame, (music concludes).
- Constraints: The host's face must remain consistent with Image 1 throughout; lip-sync must strictly match the dialogue; the studio background remains fixed and unchanged.
- General Constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural slight movements, not be like statues.
- ```
- ### No. 2: News Broadcast
- - **Category:** `Talking Head & Digital Humans`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The anchor is bound to @Image 1 (facial makeup and styling should correspond to the image), horizontal screen 16:9, authoritative news studio, cool blue professional tone, generate a 30-second news broadcast short video.
- #### Prompt
- ```text
- The anchor is bound to @Image 1 (facial makeup and styling should correspond to the image), horizontal screen 16:9, authoritative news studio, cool blue professional tone, generate a 30-second news broadcast short video.
- Shot 1: Medium shot, anchor @Image 1 sits upright, looking directly at the camera to open, blue news screen behind, fixed camera position, {Good evening viewers, welcome to this broadcast}.
- Shot 2: Slow push-in, anchor @Image 1 delivers the lead with a serious expression, steady pace, cool blue light, 【Today's Headlines】.
- Shot 3: Cut to, a news footage window appears to the right and behind anchor @Image 1, smooth pan, <studio cue sound>.
- Shot 4: Medium close-up, anchor @Image 1 turns to another camera to continue the broadcast, natural eye contact switch, fixed camera position, (low background music).
- Shot 5: Slow pull-out, anchor @Image 1 concludes with hands folded in a gesture of thanks, cool blue freeze frame, {Thank you for watching, see you next time}.
- Constraints: The anchor's face must remain consistent with Image 1 throughout; lip-sync must strictly match the dialogue; posture is upright and professional, studio background is fixed.
- General Constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural slight movements, not be like statues.
- ```
- ### No. 3: Livestream Sales Monologue
- - **Category:** `Talking Head & Digital Humans`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The host is bound to @Image 2 (facial makeup and styling should correspond to the image), vertical screen 9:16, lively livestream sales style, bright high-saturation warm tone, generate a 30-second sales monologue short video.
- #### Prompt
- ```text
- The host is bound to @Image 2 (facial makeup and styling should correspond to the image), vertical screen 9:16, lively livestream sales style, bright high-saturation warm tone, generate a 30-second sales monologue short video.
- Shot 1: Medium close-up, host @Image 2 holds up the product close to the camera, enthusiastically calling out, fixed camera position, {Family, look over here, today's product is absolutely amazing!}.
- Shot 2: Slow push-in, host @Image 2 points out the product's selling points, exaggerated excited expression, high-saturation warm light, <ding, link-up sound effect>.
- Shot 3: Cut to, a price explosion sticker appears at the bottom left of the screen, host @Image 2 claps for emphasis, smooth pan, 【Limited Time Direct Discount Only 99】.
- Shot 4: Medium shot, host @Image 2 holds up the product, showing details left and right, speaking pace quickens, fixed camera position, (energetic promotional music).
- Shot 5: Slow push-in, host @Image 2 gives a thumbs up to urge orders for the conclusion, warm tone freeze frame, {Don't miss out, hurry up and place your order!}.
- Constraints: The host's face must remain consistent with Image 2 throughout; lip-sync must strictly match the dialogue; the product's appearance remains consistent; the background shelf display is fixed.
- General Constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural slight movements, not be like statues.
- ```
- ### No. 4: Late-Night Emotional Radio
- - **Category:** `Talking Head & Digital Humans`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The host is bound to @Image 3 (facial makeup and styling should correspond to the image), vertical screen 9:16, late-night radio atmosphere, warm orange low-contrast glow, generate a 30-second emotional radio short video.
- #### Prompt
- ```text
- The host is bound to @Image 3 (facial makeup and styling should correspond to the image), vertical screen 9:16, late-night radio atmosphere, warm orange low-contrast glow, generate a 30-second emotional radio short video.
- Shot 1: Medium close-up, host @Image 3 turns slightly towards the microphone, eyes closed, speaking softly, warm orange desk lamp side lighting, fixed camera position, (gentle piano starts) {It's late, how are you who haven't slept yet?}.
- Shot 2: Slow push-in, host @Image 3 opens eyes looking into the distance with a gentle expression, eyes glistening, warm light suffusion, <sound of light night rain outside the window>.
- Shot 3: Smooth pan, the camera glides past headphones and a warm teacup on the desk, host @Image 3 enters frame whispering, (piano intensifies).
- Shot 4: Close-up, host @Image 3 smiles slightly, offering soft comfort, eyelashes flutter slightly, warm orange glow, {May you be treated gently by this world}.
- Shot 5: Slow pull-out, host @Image 3 takes off headphones, lowers head with a smile to conclude, soft glow freeze frame, (music concludes).
- Constraints: The host's face must remain consistent with Image 3 throughout; lip-sync must strictly match the dialogue; emotions are externalized through eyelash flutters and glistening eyes, avoid straightforward narration-style performance; the warm orange night tone remains consistent throughout.
- General Constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural slight movements, not be like statues.
- ```
- ### No. 5: Teaching Instructor
- - **Category:** `Talking Head & Digital Humans`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The instructor is bound to @Image 1 (facial makeup and styling should correspond to the image), horizontal screen 16:9, bright professional online classroom, clean whiteboard tone, generate a 30-second teaching instructor short video.
- #### Prompt
- ```text
- The instructor is bound to @Image 1 (facial makeup and styling should correspond to the image), horizontal screen 16:9, bright professional online classroom, clean whiteboard tone, generate a 30-second teaching instructor short video.
- Shot 1: Medium shot, instructor @Image 1 stands in front of a whiteboard, turns to face the camera to begin the lecture, holding a marker, fixed camera position, {In this lesson, we'll only cover one core formula}.
- Shot 2: Smooth pan, instructor @Image 1 walks while writing keywords on the whiteboard, camera follows the movement, bright white light, <sound of marker writing>.
- Shot 3: Cut to, the whiteboard writing area zooms in, the formula appears highlighted, instructor @Image 1 turns sideways to point, slow push-in, 【Core Formula】.
- Shot 4: Medium close-up, instructor @Image 1 turns back to the camera to explain with an example, lively expression, fixed camera position, (lighthearted learning background music).
- Shot 5: Slow pull-out, instructor @Image 1 spreads hands with a smile to conclude and nods, white light freeze frame, {Understood? See you next lesson}.
- Constraints: The instructor's face must remain consistent with Image 1 throughout; lip-sync must strictly match the dialogue; the whiteboard content is coherent from start to finish; the classroom background is fixed.
- General Constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural slight movements, not be like statues.
- ```
- ### No. 6: Multilingual Clone (Including Editing Task Concept)
- - **Category:** `Talking Head & Digital Humans`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The digital human is bound to @Image 1 (facial makeup and styling should correspond to the image), vertical screen 9:16, unified studio style for a multinational brand, simple tech blue tone, generate a 30-second multilingual clone short video.
- #### Prompt
- ```text
- The digital human is bound to @Image 1 (facial makeup and styling should correspond to the image), vertical screen 9:16, unified studio style for a multinational brand, simple tech blue tone, generate a 30-second multilingual clone short video.
- Shot 1: Medium close-up, digital human @Image 1 looks directly at the camera, opening with a greeting in Chinese, tech blue background screen, fixed camera position, {Hello everyone, welcome to learn about our product}【Chinese】.
- Shot 2: Cut to, the same digital human @Image 1 seamlessly switches to English monologue, lip-sync aligned to English, slow push-in, {Hello everyone, welcome}【English】.
- Shot 3: Cut to, digital human @Image 1 switches to Japanese greeting, consistent demeanor, smooth pan, {Mina-san konnichiwa}【Japanese】.
- Shot 4: Medium shot, digital human @Image 1 opens arms in a welcoming gesture, multilingual subtitle stream appears in the background, fixed camera position, (tech-sounding background music).
- Shot 5: Slow pull-out, digital human @Image 1 smiles and nods to conclude, blue tone freeze frame, (music concludes).
- Constraints: Editing task concept — First generate the Chinese master shots, then use the editing instruction "keep the character and camera movements unchanged, only replace the dialogue and lip-sync with the target language" to derive the English/Japanese clones, ensuring the face @Image 1, camera positions, background, and attire remain completely consistent throughout; lip-sync for each language must strictly match its corresponding dialogue; prevent facial drift in the clones.
- General Constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural slight movements, not be like statues.
- ```
- <a id="category-16"></a>
- ### Short Dramas & Emotional Stories (8)
- ### No. 1: Convenience Store Reunion (Emotional Reversal)
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Refer to the female lead in Image 1 and the male lead in Image 3 (their facial makeup and styling should match the corresponding images), vertical 9:16, cinematic realism, cold blue transitioning to warm orange, generate a 30-second emotional short drama.
- #### Prompt
- ```text
- Refer to the female lead in Image 1 and the male lead in Image 3 (their facial makeup and styling should match the corresponding images), vertical 9:16, cinematic realism, cold blue transitioning to warm orange, generate a 30-second emotional short drama.
- Shot 1: Wide shot, the female lead @Image 1 leans against the window alone, head bowed, stirring her coffee, cold blue night tone, the camera slowly pushes in, (light piano begins).
- Shot 2: Medium close-up, the male lead @Image 3 pushes the door open and enters the frame, the female lead @Image 1 looks up and freezes, fixed camera with a slight breathing feel, <doorbell dings>.
- Shot 3: Close-up, the shot cuts back and forth between the two, the color tone gradually warms, the male lead @Image 3 says softly {I'm here to pick you up}, (piano intensifies).
- Shot 4: Close-up, the female lead @Image 1's eyes well up with tears and the corners of her mouth lift into a smile, the camera slowly pulls out, warm orange freeze frame, (music resolves).
- Constraints: The faces of the male and female leads must remain consistent with the corresponding images throughout; the convenience store setting is fixed; the emotional reversal is externalized through teary eyes and trembling lips; do not generate clones/twins.
- General constraints: Keep it without subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be like statues.
- ```
- ### No. 2: Rainy Night Monologue
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The protagonist is bound to @Image 2 (facial makeup and styling should match the corresponding image), vertical 9:16, cool-toned artistic realism, bluish-grey rainy night with low saturation, generate a 30-second emotional monologue short film.
- #### Prompt
- ```text
- The protagonist is bound to @Image 2 (facial makeup and styling should match the corresponding image), vertical 9:16, cool-toned artistic realism, bluish-grey rainy night with low saturation, generate a 30-second emotional monologue short film.
- Shot 1: Close-up, rainwater trickles down the glass window, the protagonist's @Image 2 blurred reflection appears behind the glass, fixed camera, <sound of light rain> (low cello begins).
- Shot 2: Slow push-in, the protagonist @Image 2 presses against the window gazing at the streetlights in the rainy night, fingertips touching the cold glass, bluish-grey cold light, {Some words can only be told to the rain}.
- Shot 3: Smooth pan, the camera glides past a phone with its screen off and a half-cold cup of tea on the table, the protagonist's @Image 2 profile enters the frame, (cello intensifies).
- Shot 4: Medium close-up, the protagonist @Image 2 takes a deep breath and closes their eyes, a single tear traces down their cheek, overlapping with a raindrop on the window, slow push-in, <rain intensifies>.
- Shot 5: Slow pull-out, the protagonist @Image 2 opens their eyes, pulls a relieved faint smile, and turns away from the window, bluish-grey freeze frame, (music resolves).
- Constraints: The protagonist's face must remain consistent with Image 2 throughout; emotions are externalized through fingertips against the glass and tear tracks, avoid direct dialogue like "I'm very sad"; the bluish-grey rainy night tone and setting remain consistent throughout.
- General constraints: Keep it without subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be like statues.
- ```
- ### No. 3: Family Dining Table
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Refer to the mother in Image 1, the father in Image 2, and the child in Image 4 (their facial makeup and styling should match the corresponding images), horizontal 16:9, warm life realism, warm yellow home tone, generate a 30-second family emotional short film.
- #### Prompt
- ```text
- Refer to the mother in Image 1, the father in Image 2, and the child in Image 4 (their facial makeup and styling should match the corresponding images), horizontal 16:9, warm life realism, warm yellow home tone, generate a 30-second family emotional short film.
- Shot 1: Wide shot, the family of three sits around the dining table, steam rises from hot dishes, the mother @Image 1 is serving food to the child @Image 4, warm yellow lighting, fixed camera, (warm folk music begins) <clinking of bowls and chopsticks>.
- Shot 2: Medium close-up, the father @Image 2 smiles as he serves soup to the mother @Image 1, they exchange a smile, slow push-in, <clinking of soup ladle>.
- Shot 3: Close-up, the child @Image 4 puffs their cheeks while eating heartily, looks up and grins, warm light falls on their face, cut to this shot.
- Shot 4: Medium shot, the mother @Image 1 reaches out to wipe a corner of the father's @Image 2 mouth, the father is momentarily surprised then holds her hand, smooth pan, (folk music becomes warmer).
- Shot 5: Slow crane up, the camera rises to look down on the table full of food and the family leaning close together, warm yellow freeze frame, (music resolves).
- Constraints: The faces of the three characters must strictly match the corresponding images, the mother/father/child should have distinct body types and ages, do not generate clones/twins or facial confusion; the dishes on the table and the warm yellow home setting remain consistent throughout.
- General constraints: Keep it without subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be like statues.
- ```
- ### No. 4: Farewell at the Station
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Refer to the female lead in Image 1 and the male lead in Image 3 (their facial makeup and styling should match the corresponding images), horizontal 16:9, cinematic realism, cool-toned platform with slightly warm backlight, generate a 30-second farewell emotional short film.
- #### Prompt
- ```text
- Refer to the female lead in Image 1 and the male lead in Image 3 (their facial makeup and styling should match the corresponding images), horizontal 16:9, cinematic realism, cool-toned platform with slightly warm backlight, generate a 30-second farewell emotional short film.
- Shot 1: Wide shot, the platform is crowded with people, the female lead @Image 1 and the male lead @Image 3 stand facing each other, holding hands tightly, a train is beside them, cool tone with backlight, slow push-in, (low strings begin) <train arrival announcement>.
- Shot 2: Close-up, their tightly clasped fingers slowly tighten then reluctantly loosen, warm backlight outlines them, fixed camera.
- Shot 3: Medium close-up, the male lead @Image 3 reaches up to tuck a strand of wind-blown hair behind the female lead's @Image 1 ear, hesitates to speak, slow push-in, {When you arrive... remember to let me know you're safe}.
- Shot 4: Medium shot, the male lead @Image 3 turns and boards the train, the female lead @Image 1 stands still, her shoulders trembling slightly as she holds back tears, smooth pan, (strings rise) <door closing warning sound>.
- Shot 5: Slow pull-out, the train departs, the female lead @Image 1 stands alone on the empty platform as a backlit silhouette, cool tone freeze frame, (music resolves).
- Constraints: The faces of the male and female leads must remain consistent with the corresponding images throughout; the farewell emotion is externalized through tightening fingers and trembling shoulders, avoid direct crying and shouting; the station platform setting and the interplay of cool/warm backlight remain consistent throughout; do not generate clones/twins.
- General constraints: Keep it without subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be like statues.
- ```
- ### No. 5: Workplace Highlight
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The protagonist is bound to @Image 2 (facial makeup and styling should match the corresponding image), horizontal 16:9, business-inspiring realism, cold blue steel/glass texture transitioning to golden light, generate a 30-second workplace motivational short film.
- #### Prompt
- ```text
- The protagonist is bound to @Image 2 (facial makeup and styling should match the corresponding image), horizontal 16:9, business-inspiring realism, cold blue steel/glass texture transitioning to golden light, generate a 30-second workplace motivational short film.
- Shot 1: Medium close-up, the protagonist @Image 2 sits alone in a late-night office, frowning at a screen full of data, cold blue screen light illuminates their face, fixed camera, (low drumbeat begins) <intense keyboard typing>.
- Shot 2: Slow push-in, the protagonist @Image 2 suddenly looks up, eyes lighting up, grabs a pen and circles a key point, the screen light brightens sharply, <inspiration strike sound effect>.
- Shot 3: Follow shot, the protagonist @Image 2 stands up and strides through the hallway, pushing open the conference room door, the camera follows closely behind their back, (drumbeat intensifies).
- Shot 4: Low-angle shot, the protagonist @Image 2 stands before a floor-to-ceiling window, confidently presenting to a group, city skyline in the background, backlit with golden light, {This is our answer}.
- Shot 5: Slow orbit, the entire room stands up applauding, the protagonist @Image 2 lets out a sigh of relief and smiles, golden light freeze frame, (music climax resolves) <thunderous applause>.
- Constraints: The protagonist's face must remain consistent with Image 2 throughout; the emotional arc is externalized from frowning to relaxation, the transition from cold blue to golden light corresponds to low point to highlight; the office setting is coherent and consistent.
- General constraints: Keep it without subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be like statues.
- ```
- ### No. 6: Campus Youth
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Refer to the girl in Image 4 and the boy in Image 5 (their facial makeup and styling should match the corresponding images), vertical 9:16, youthful fresh realism, bright warm sunlight with high translucency, generate a 30-second campus youth short film.
- #### Prompt
- ```text
- Refer to the girl in Image 4 and the boy in Image 5 (their facial makeup and styling should match the corresponding images), vertical 9:16, youthful fresh realism, bright warm sunlight with high translucency, generate a 30-second campus youth short film.
- Shot 1: Wide shot, warm afternoon sunlight slants across the sports field, the girl @Image 4 walks quickly across the track holding books, the wind lifts her skirt, slow push-in, (fresh guitar begins) <light cicada chirping>.
- Shot 2: Medium close-up, the boy @Image 5 rides past on a bicycle, brakes suddenly and looks back, their gazes meet and they both freeze for a moment, fixed camera, <bicycle brake sound>.
- Shot 3: Close-up, the girl's @Image 4 cheeks flush pink, she lowers her head and smiles with pursed lips, the scattered books are picked up by the boy @Image 5 as he bends down, cut to this shot in warm sunlight.
- Shot 4: Medium shot, the two walk side by side along a tree-lined path talking and laughing, light spots dance on their faces, follow shot accompanying them, (guitar becomes more cheerful).
- Shot 5: Slow crane up, the camera rises to look down on their receding backs and the campus bathed in golden light, warm sunlight freeze frame, (music resolves).
- Constraints: The faces of the girl and boy must remain consistent with the corresponding images throughout, their male/female body types should be distinct, do not generate clones/twins or confusion; the campus setting with bright warm sunlight and high translucency remains consistent throughout; youthful emotions are externalized through flushed cheeks and pursed-lip smiles.
- General constraints: Keep it without subtitles; do not generate a Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural, subtle movements, not be like statues.
- ```
- ### No. 7: Suspenseful Corridor
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- The protagonist is bound to @Image 2 (facial makeup and styling should correspond to the reference image), vertical screen 9:16, thriller suspense realistic style, dim cold green with high-contrast shadows, generate a 30-second suspense short film.
- #### Prompt
- ```text
- The protagonist is bound to @Image 2 (facial makeup and styling should correspond to the reference image), vertical screen 9:16, thriller suspense realistic style, dim cold green with high-contrast shadows, generate a 30-second suspense short film.
- Shot 1: Close-up, a corridor light flickers on and off, the protagonist @Image 2's flashlight beam pierces the darkness, dim cold green, fixed camera, <buzzing current sound from the light tube> (low-frequency hum begins).
- Shot 2: Following shot, the camera closely follows behind the protagonist @Image 2 as they slowly walk deeper into the long corridor, hesitant footsteps, elongated shadows, <echoing footsteps>.
- Shot 3: Slow push-in, the protagonist @Image 2 suddenly stops, holds their breath, and looks back, pupils contract, sweat beads form, cold green light illuminates their face, <strange clicking sound from behind>.
- Shot 4: Cut to, a door at the end of the corridor slowly opens with a sliver of light seeping out, the protagonist @Image 2 raises their light and steps back cautiously, (sharp strings rise sharply).
- Shot 5: Slow pull-out, the protagonist @Image 2 stands alone in the center of the long, dark corridor, shadows closing in from all sides, cold green freeze-frame, <sudden silence> (music cuts off abruptly).
- Constraints: The protagonist's face must remain consistent with Image 2 throughout; tense emotions are externalized through holding breath, looking back, and sweating with contracted pupils; dim cold green high-contrast lighting and corridor setting remain consistent throughout; use silence at the end to create a suspenseful open ending.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural subtle movements, not be like statues.
- ```
- ### No. 8: Reconciliation (Using Editing Task for Contrastive Derivation)
- - **Category:** `Short Dramas & Emotional Stories`
- - **Source:** `community`
- - **Author:** Atlas Cloud
- - **Language:** `en`
- #### Description
- Refer to the female lead from Image 1 and the male lead from Image 3 (their respective facial makeup and styling should correspond to the reference images), vertical screen 9:16, cinematic realism, transitioning from cold gray rainy to warm golden sunny light, generate a 30-second emotional short film about reconciliation.
- #### Prompt
- ```text
- Refer to the female lead from Image 1 and the male lead from Image 3 (their respective facial makeup and styling should correspond to the reference images), vertical screen 9:16, cinematic realism, transitioning from cold gray rainy to warm golden sunny light, generate a 30-second emotional short film about reconciliation.
- Shot 1: Medium close-up, the female lead @Image 1 and the male lead @Image 3 stand back-to-back in the cold gray rain, heads bowed, distance between them, fixed camera, (oppressive piano begins) <cold, desolate sound of rain>.
- Shot 2: Close-up, an old photo lies spread out in a puddle, getting wet from the rain, in the mirror-like reflection the two grow further apart, slow push-in.
- Shot 3: Cut to, the sky clears, warm golden light falls, the male lead @Image 3 turns and reaches out a hand, the female lead @Image 1 slowly looks back with tears in her eyes, smooth lateral movement, (piano transitions to warm and intensifies).
- Shot 4: Medium shot, the two embrace, the female lead @Image 1 buries her head in the male lead @Image 3's shoulder, the male lead gently strokes her hair, warm golden backlight outlines them, slow circling, {I'm sorry, I'm back}.
- Shot 5: Slow pull-out, the two figures leaning on each other merge into one under the warm golden sunlight, freeze-frame, (music concludes satisfactorily).
- Constraints: Editing task approach—first generate shots 1 to 2 as the "cold gray rainy estrangement version" master, then use the editing instruction "keep the faces of @Image 1/@Image 3 and the camera composition unchanged, only change the weather to warm golden sunlight, and the posture from back-to-back to embracing" to derive the contrasting version shots 3 to 5, achieving a warm/cold reversal in the same scene; the faces of the male and female leads must remain consistent throughout, no doppelgangers/twins; reconciliation emotions are externalized through looking back with tears and embracing/burying head.
- General constraints: Keep no subtitles; do not generate Logo; do not generate watermarks; use only one camera movement per shot; characters should have natural subtle movements, not be like statues.
- ```
- <a id="category-17"></a>
Lines 2,001–2,563 are not shown. The link to the file on GitHub has every line.
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.