
Seedance 2.5 is designed less like a one-shot text-to-video toy and more like a compact production workflow. You can describe a scene, assign jobs to image, video, and audio references, plan a 30-second sequence, refine an existing clip, or use rough visual blocking to guide a polished result.

This guide focuses on the practical part: how to turn an idea and a pile of assets into instructions the model can follow.

## The Core Prompt Formula

A dependable Seedance 2.5 prompt answers seven questions:

> **Subject + environment + action + camera + look + audio + constraints**

- **Subject:** Who or what must remain recognizable?
- **Environment:** Where and when does the scene happen?
- **Action:** What changes from the first moment to the last?
- **Camera:** What is the framing, lens feel, movement, and edit rhythm?
- **Look:** What lighting, color, texture, and medium define the image?
- **Audio:** What dialogue, music, ambience, and effects belong on the timeline?
- **Constraints:** What must stay fixed, and what must not appear?

Write the main creative intent first. Add references and timing second. Put non-negotiable continuity rules last. A prompt should read like a concise director's brief, not a bag of visual adjectives.

## Map Every Reference to a Role

Uploading more material does not automatically produce more control. The model needs to know what each asset contributes.

Use explicit assignments such as:

- `@Image 1` defines the lead character's face, hair, proportions, and wardrobe.
- `@Image 2` defines the location, palette, and lighting only.
- `@Video 1` defines choreography and timing, but not identity or clothing.
- `@Video 2` defines camera movement only.
- `@Audio 1` defines the musical beat and overall duration.

If two references disagree, state which one wins. For example: “Use the person from `@Image 1`; borrow only the body motion from `@Video 1`; ignore the performer and background in that video.” This prevents a motion reference from quietly replacing your cast or art direction.

### Choose references scene by scene

Do not spend the entire reference budget on every shot. Select the smallest useful set for each scene:

- **Character dialogue:** identity sheet, wardrobe frame, location frame, and voice or dialogue reference.
- **Dance or action:** character sheet, motion video, camera-motion video if needed, and rhythm track.
- **Product commercial:** product packshots from useful angles, material close-ups, set reference, and a camera or lighting reference.
- **Storyboard sequence:** numbered boards or keyframes, character lock, and one consistent style frame.
- **Video edit:** the source clip plus only the references required for the requested replacement or restyle.

For multi-character work, name each person and reserve a distinct reference group for them. Repeat the mapping near the timeline whenever a character returns after several shots.

## Dreamina Reference Limits and Prompt Syntax

The Dreamina guide describes a **maximum of 50 references in total**, with these category limits:

- Up to **30 images**, each input image no larger than **4K**.
- Up to **10 videos**, with all reference video duration totaling no more than **30 seconds**.
- Up to **10 audio files**, with all reference audio duration totaling no more than **30 seconds**.

The 4K figure above is an **input-image limit**, not a promise of 4K video output.

Dreamina also documents lightweight syntax for separating sound and text instructions:

- Music: `(soft analog synth pulse, gradually building)`
- Sound effect: `<ceramic cup placed on a wooden counter>`
- Dialogue: `{I knew you would come back.}`
- Subtitle or on-screen caption: `【Three hours earlier】`

For non-Chinese dialogue, put the language before the line. You can also specify a regional variety, accent, and delivery, for example: `English, Irish accent, quietly amused: {You took your time.}`

Treat these markers as clear labels, not magic commands. The surrounding scene, speaker, timing, and emotional direction still matter.

## Write 30 Seconds as Timed Beats

Thirty seconds is long enough for a scene to drift if the prompt contains only one action. Divide it into a small number of meaningful phases. Four to six beats are usually easier to direct than thirty one-second instructions.

A useful dramatic shape is:

1. **0–5s — Establish:** location, subject, and visual question.
2. **5–12s — Develop:** the main action begins and gains direction.
3. **12–20s — Escalate:** a reveal, complication, or change in camera energy.
4. **20–27s — Payoff:** the decisive action or emotional turn.
5. **27–30s — Resolve:** a readable final image with room to land.

Each beat should specify what changes. “The camera moves” is vague; “a slow dolly-in changes a waist-up two-shot into a tight close-up while she realizes the letter is hers” gives the movement a purpose.

Keep transitions causal. The subject's final position, gaze, momentum, lighting, and screen direction in one phase should give the next phase somewhere logical to begin.

## Lock Parameters Before Directing Motion

References establish possibilities; locks establish priorities. Put a compact lock block before the timeline when consistency matters.

Lock only what must not drift:

- Character identity, age range, body proportions, hair, and wardrobe.
- Product shape, label placement, materials, and brand colors.
- Location layout, time of day, weather, and primary light direction.
- Aspect ratio, visual medium, color treatment, and camera axis.
- Required props, character count, and spoken language.

Then separate controlled variation from fixed parameters: “Wardrobe and face remain unchanged; expressions progress from guarded to relieved.” This tells the model that emotional change is intentional rather than identity drift.

Do not issue incompatible locks. A fixed focal length conflicts with a requested zoom-lens effect; a locked noon sun conflicts with a sunset transition. Decide which parameter carries the story.

## Direct Observable Emotion

Words such as “sad” or “nervous” name an internal state but do not tell the camera what to see. Convert emotion into behavior:

- Her smile arrives late and disappears when she notices the empty chair.
- He rehearses the greeting under his breath, wipes his palm on his coat, then knocks.
- She avoids eye contact, inhales once, and finally holds his gaze on the last line.
- His shoulders loosen only after the off-screen door closes.

Pair one facial cue with posture, breath, gaze, or hand behavior. For dialogue, add vocal delivery and a reaction beat after the line. Small, timed actions usually read more convincingly than a demand for “extreme emotion.”

## Use Professional Camera Language Precisely

Camera terms are useful when they describe a visible production choice:

- **Shot size:** extreme wide, wide, medium, medium close-up, close-up, extreme close-up.
- **Angle:** eye level, low angle, high angle, overhead, over-the-shoulder, Dutch angle.
- **Movement:** pan, tilt, dolly in or out, truck left or right, crane, pedestal, orbit, handheld follow.
- **Lens behavior:** wide-angle spatial depth, natural perspective, compressed telephoto background, shallow depth of field, rack focus.
- **Staging:** two-shot, profile composition, foreground occlusion, deep staging, clean single, shot-reverse-shot.
- **Editing:** hard cut, match cut, whip-pan transition, motivated cut, continuous take, J-cut, L-cut.

Do not stack movements simply because they sound cinematic. “A slow lateral track as the cyclist accelerates” is clearer than “pan, orbit, crane, zoom, and handheld.” Define the subject-camera relationship and one dominant move per beat.

## First and Last Frames, Keyframes, and Storyboards

Different planning assets solve different problems.

### First and last frames

Use a first frame to lock the opening composition. Add a last frame when the destination matters just as much: a product must end centered, a transformation must reach a specific design, or a character must arrive at an exact pose.

Describe the path between them. Without an action bridge, the model may morph from one state to the other rather than stage a physical transition.

### Multiple keyframes

Use keyframes when a shot has several required visual milestones. Number them in order, assign approximate times, and state whether the camera cuts or moves continuously between them. Keep neighboring frames compatible in character, geography, and lighting.

### Storyboards

A storyboard controls shot order and composition. Tell the model whether each panel represents a cut, a camera destination, or an action peak. Do not assume visible panel numbers or annotations should appear in the final video; explicitly exclude board borders, labels, arrows, and production notes.

### Rough and refined blockouts

A rough blockout—simple shapes, an animatic, a 3D white model, or green-screen staging—is useful for timing, spatial layout, and camera paths. A refined blockout adds clearer poses, lens choices, lighting direction, and edit points. In both cases, state what should be preserved and what should be replaced by the final style.

The official ByteDance Seed page specifically identifies white-model control and green-screen editing as Seedance 2.5 production capabilities.

## One-Click Video Without Losing Control

For a fast one-click workflow, prepare one compact production packet:

1. A clear prompt using the core formula.
2. A small, role-labeled reference set.
3. A lock block for identity, wardrobe, setting, and format.
4. A timed sequence for anything longer than a single simple action.
5. An end-state and a short exclusion list.

Then generate the full shot in one pass. “One click” works best when the planning has already removed ambiguity. If the first result misses one isolated element, edit that element rather than rewriting the entire concept.

## Edit Existing Video with Surgical Instructions

Seedance 2.5 can be directed to modify existing video. Write edit prompts in four parts:

> **Keep + change + affected time or region + continuity requirement**

Example:

> Keep the source camera movement, actor performance, timing, shadows, and ambient audio. From 06–11 seconds, replace only the paper cup in her right hand with the blue ceramic mug from `@Image 1`. Preserve hand contact, reflections, scale, occlusion, and motion blur. Do not alter her face, wardrobe, background, or the rest of the timeline.

For green-screen material, state how the new environment should match the subject: contact shadows, color spill, depth of field, perspective, and interactive light. For white-model footage, identify whether the blockout controls staging, motion, camera, or all three.

## Extend Forward or Backward with Continuity

ByteDance's official page says a generated video can be extended **twice**. An extension should begin from a continuity handoff, not a new synopsis.

For a forward extension, restate the final observable state:

- Subject position, facing direction, pose, and motion vector.
- Camera height, framing, lens feel, and movement.
- Lighting direction, weather, and background activity.
- Wardrobe, props, damage, and other accumulated changes.
- Audio ambience, musical phase, and unfinished dialogue.

For a backward extension, define the required opening state of the existing clip and reverse-engineer a plausible lead-in. The new segment must arrive at the original first frame with matching momentum, exposure, subject placement, and sound.

Avoid repeating the same action at the seam. Ask for continuation from the exact contact, step, glance, or camera movement already underway. If possible, use the boundary frame and the source clip together.

## Design Seamless Transitions

A seamless transition needs a shared visual or physical bridge. Reliable options include:

- **Match on action:** a hand movement continues across the cut.
- **Shape or color match:** a circular lamp becomes the moon; a red fabric fills the frame and reveals a new location.
- **Occlusion:** a foreground object wipes across the lens.
- **Whip pan:** directional blur carries the same speed and direction into the next scene.
- **Lighting bridge:** a flash, shadow, or exposure change motivates the transformation.
- **Sound bridge:** the next scene's audio begins before the visual cut, or the previous sound continues after it.

Specify the seam's direction, speed, framing, and audio. “Transition seamlessly” alone leaves too many choices open.

## Copy-Ready Prompt Templates

Replace the bracketed fields and reference numbers with your own material.

### 30-second cinematic scene

```text
Create a 30-second [genre] scene in [aspect ratio].

REFERENCE MAP
@Image 1: lock [character] identity, hair, proportions, and wardrobe.
@Image 2: use only for [location], palette, and light direction.
@Video 1: use only for [movement/choreography], not identity or setting.
@Audio 1: use as the timing and musical reference.

LOCKS
Keep [identity/product/location parameters] unchanged throughout. Maintain [camera axis, lighting, style, character count].

TIMELINE
0–5s: [establishing composition and first action].
5–12s: [development, camera relationship, observable emotion].
12–20s: [escalation or reveal].
20–27s: [payoff action and strongest camera beat].
27–30s: [resolved final image and audio landing].

AUDIO
(music direction)
<key synchronized sound effect>
[Language, accent, delivery]: {dialogue}

Exclude [only the most important unwanted changes, artifacts, text, logos, or extra subjects].
```

### Product video from multiple references

```text
Create a clean 15-second premium product film for the object in @Image 1.
Use @Image 2 for the side profile and @Image 3 for material detail only. Preserve the exact silhouette, proportions, finish, controls, and logo placement from every angle. Use @Image 4 only for the warm studio palette.

0–4s: Macro detail of [feature], slow lateral track, narrow highlight moving across the surface.
4–10s: Pull back into a three-quarter hero view as [physical product action] occurs with realistic weight and contact.
10–15s: Controlled half-orbit ending on the exact front view from @Image 1, centered with negative space for copy to be added in post.

<precise product sound>, (restrained minimal score). No generated captions, extra controls, warped branding, floating parts, or shape changes.
```

### Emotion-led dialogue scene

```text
Two characters wait on an empty late-night train platform. Lock the identities and wardrobes from @Image 1 and @Image 2. Cool overhead light, soft rain beyond the platform roof, naturalistic drama, restrained handheld camera.

Begin in a quiet waist-up two-shot. Character A grips a folded ticket, tries to speak, then looks away. Character B notices the gesture but does not interrupt. Slowly push into Character A as their shoulders settle.

English, London accent, barely above a whisper: {I kept the ticket.}
<distant train brakes, rain on metal roofing>
(very soft low strings entering only after the line)

Hold the reaction for two seconds: Character B exhales, looks at the ticket, and gives a small disbelieving smile. Keep eyelines, rain direction, background geography, and facial identity consistent. No subtitles.
```

### Precise source-video edit

```text
Edit @Video 1 without changing its duration, shot order, camera motion, performance, or audio timing.

Change only [object/person/region] during [time range]. Replace it with [description or @Image reference]. Match the source perspective, scale, lighting, reflections, motion blur, occlusion, and physical contact in every affected frame.

Keep [protected elements] exactly as in the source. Do not modify any frame outside the stated range. No new text, logos, props, or people.
```

### Forward extension

```text
Extend @Video 1 forward from its exact final frame. Begin with [subject] at [position and pose], moving [direction and speed], while the camera continues its [existing movement]. Match the existing lens feel, exposure, light direction, weather, wardrobe, props, environment, ambience, and music phase.

Continue the unfinished action: [next causal action]. Do not replay the ending or pause at the seam. Over [duration], develop toward [new payoff], then finish on [specific final composition]. Preserve identity and spatial geography throughout.
```

## Pre-Flight Checklist

Before generating, confirm:

- The prompt has one clear creative goal.
- Every reference has one named role, and conflicts have a priority rule.
- Character, product, setting, and format locks are explicit but not contradictory.
- A long clip has timed beats with visible changes.
- Actions connect physically across beats, cuts, and extensions.
- Camera terms describe one dominant choice per moment.
- Emotion is expressed through observable behavior and vocal delivery.
- Dialogue, music, sound effects, and captions use clear labels and timing.
- The final state is defined.
- Exclusions are short and directly relevant.
- Reference counts and combined durations stay within the Dreamina guide's stated limits.
- You have not treated the per-image 4K input ceiling as an output-resolution claim.
- You have permission to use the people, brands, footage, voices, and music in your references.

## Limitations and Evidence Boundary

Seedance 2.5 is generative, so prompt compliance, identity, readable text, anatomy, object permanence, dialogue, and audio synchronization can still vary between results. Review every output before publishing, especially for brand assets, recognizable people, product claims, safety-sensitive actions, and licensed material.

The official ByteDance Seed page confirms **single generations up to 30 seconds**, **up to two extensions**, more precise reference understanding and editing, and production workflows involving **white models and green screens**. It does **not** publicly specify 4K output, API availability, pricing, regional availability, generation speed, or commercial licensing terms on that page. Do not infer those details from the 4K input-image limit in the Dreamina guide. Product interfaces and limits may also change after publication, so verify current settings in the service you use.

## Put the Workflow Into Practice

Explore more ready-made structures, or start creating with Seedance 2.5 in SJinn.

[Browse Seedance 2.5 prompt examples](/seedance-2-5-prompts)

[Create a Seedance 2.5 video](/tool-mode/seedance25-video)


Learn how to direct Seedance 2.5 with structured prompts, multimodal references, timed scenes, precise edits, extensions, keyframes, and production-ready templates.

The Complete Guide to Seedance 2.5: Prompts, References, Editing, and 30-Second Storytelling


# Seedance 2.0 Prompt Collection

> 精选高信噪比 AI 视频提示词，持续更新中 — 共计 203 条提示词

Seedance 2.0 是字节跳动最新的 AI 视频生成模型，可生成长达 **8 秒**、原生 **1080p** 分辨率的影视级画质视频。下面是从社区精选的优秀提示词集合，按场景类型分类整理，帮助你快速上手。

---

## Featured Picks / 精选推荐

以下三条提示词是社区评选的最佳范例，涵盖了 Seedance 2.0 最强大的能力维度。

---

### 1. 超现实战场浪人动作场景

**类型:** 多镜头 / 动作 / 超现实

一个高度详细、多镜头提示，描述了一个超现实、高动作的空中战场场景。一名蒙面浪人被一只胸部是暴风云漩涡的巨大有翼野兽追逐。该提示包括具体的摄像机运动、技术细节（720p、16:9、15 秒）以及戏剧性的结局。

**亮点:**
- 多镜头叙事结构
- 精确的摄像机运动指令
- 超现实生物设计描述
- 完整的技术参数规格

---

### 2. 擎天柱大战哥斯拉在东京

**类型:** CGI / 科幻 / 城市战斗

Seedance 2.0 生成的 CGI 视频：一辆豪华轿车在东京的雨夜背景下，变形为擎天柱，与哥斯拉展开激战，伴随着爆炸和能量冲击波。

**亮点:**
- IP 角色融合（变形金刚 × 哥斯拉）
- 复杂场景物理模拟（雨夜、爆炸、冲击波）
- 动态变形过程描述

---

### 3.《鬼灭之刃》真人战斗提示词

**类型:** 真人改编 / 动漫 / 战斗

一个为 Seedance 2.0 设计的详细、高能量视频提示，生成 15 秒《鬼灭之刃》风格战斗的真人改编片段 — 水之呼吸 vs. 雷之呼吸。

**详细规格:**

| 参数 | 设置 |
|------|------|
| 风格 | 好莱坞真人漫画改编、黑暗武士 |
| 画质 | 4K、极致快速剪辑、粒子光效 |
| 场景 | 夜晚迷雾森林 |
| 镜头 | 三个不同视角镜头 |

包含角色动作、能力提升序列和最终冲突的完整描绘。

---

## Action & Battle / 动作与战斗

这一类别聚焦于战斗场景、动作序列和武打编排。Seedance 2.0 在复杂物理模拟和多主体交互方面表现出色。

---

### 多场景电影级战斗（孙悟空大战祖国人）

**类型:** 多场景 / 超级英雄 / 跨文化

一个详细的多场景提示，生成孙悟空与祖国人（The Boys）之间电影般的战斗序列。为每个 4 秒场景指定了摄像机运动、氛围、动作和视觉效果，并配有音乐提示。

**技巧点:** 将每个场景控制在 4 秒，分别指定镜头语言，是多镜头叙事的最佳实践。

---

### 吴宇森风格枪战场景

**类型:** 枪战 / 电影致敬 / 风格化

使用 Seedance 2.0 生成吴宇森标志性的狭窄空间枪战场景，强调电影节奏、不同摄像机角度以及特定角色参考图像的使用。

**技巧点:** 通过引用特定导演风格（如"吴宇森风格"），Seedance 2.0 可以理解并模拟其标志性的慢动作双枪美学。

---

### 武侠剑客对决

**类型:** 武侠 / 电影化 / 慢动作

一个高度详细、电影化的提示，描述了一部武侠风格的视听大片。两名剑客在竹林中决斗，重点关注镜头运动、氛围以及打斗中的超慢动作效果。

**技巧点:** 竹林 + 剑客 + 超慢动作是经过验证的高质量组合，物理模拟效果极佳。

---

### 特工电影打斗场景

**类型:** 间谍 / 动作 / 幽默

生成一个间谍电影打斗场景：一个穿绿色西装的男人和一个穿黑色衣服的男人之间的对决，结局带有幽默反转。

**技巧点:** 在动作场景中加入幽默元素，可以让生成结果更具故事性和观赏性。

---

## Narrative & Storytelling / 叙事与故事

---

### 疯狂的麦克斯：出埃及记故事

**类型:** 史诗 / 多镜头 / 叙事

一个为 Seedance 2.0 设计的提示，制作多镜头、高动作的史诗视频，讲述一个出埃及记的故事，风格类似于《疯狂的麦克斯》系列。

**技巧点:** 将经典叙事与特定电影风格结合，是 Seedance 2.0 长视频生成的有效策略。

---

### 古装刑场复仇场景

**类型:** 古装 / 戏剧 / 文字转视频

一个详细的文本转视频提示，生成设定在古代刑场的电影场景。重点突出雪、鲜血和极度严寒的视觉元素，适合戏剧性的复仇叙事。

**技巧点:** 强烈的视觉对比（白雪 × 鲜血 × 寒冷）可以充分发挥 Seedance 2.0 的环境渲染能力。

---

## Character & Dialogue / 角色与对话

---

### 美少女战士对话场景

**类型:** 动漫 / 对话 / 多角色

一个详细的基于对话的提示，生成包含水兵月、水兵金星和水兵木星的场景，包括具体的动作和对话台词。

**技巧点:** Seedance 2.0 支持多主体一致性，在对话场景中可以维持角色外观的稳定性。

---

### 卡玛拉 vs. 特朗普：提示之战

**类型:** 创意 / 动作 / 人物

尝试使用 Seedance 2.0 生成人物之间的打斗场景，可能改编自最初为动作角色设计的提示模板。

**技巧点:** 复用成熟的动作场景提示模板，替换其中的角色，是快速创作的高效方法。

---

## Creative & Lifestyle / 创意与生活

---

### 甜品店里的 360 度全景自拍

**类型:** 生活 / 自拍 / 多场景

生成一个动态视频：在一家甜品店里用 360 度全景相机自拍，随后一名女性在不同场景和服装中摆姿势。

**技巧点:** 360 度全景视角 + 场景切换 + 服装变换，是展示 Seedance 2.0 连续性和创意能力的绝佳案例。

---

## Prompt Writing Best Practices / 提示词编写最佳实践

掌握以下技巧，让你的 Seedance 2.0 提示词从 "能用" 到 "好用"。

### 核心公式

> **主体 + 动作 + 场景 + 光照 + 镜头语言 + 风格 + 画质 + 约束条件**

### 五个关键原则

| 原则 | 说明 | 示例 |
|------|------|------|
| **场景优先** | 从清晰的场景描述开始 | "夜晚迷雾森林中..." |
| **镜头明确** | 指定具体的镜头运动 | "缓慢推进"、"半环绕" |
| **光照氛围** | 定义光照和情绪氛围 | "温暖的金色夕阳光" |
| **主体详细** | 添加主体的细节描述 | "蒙面浪人，黑色斗篷飘动" |
| **控制篇幅** | 提示词控制在 200 字以内 | 精炼 > 冗长 |

### 物理模拟加分项

提及物理交互可以充分发挥模型优势：

- "风吹动头发" — 触发流体动力学模拟
- "水花飞溅" — 触发粒子系统
- "布料随风飘动" — 触发布料垂坠模拟
- "雨滴打在玻璃上" — 触发环境物理

---

## FAQ / 常见问题

### 什么是 Seedance 2.0？

Seedance 2.0 是字节跳动最新的 AI 视频生成模型，可生成长达 8 秒、原生 1080p 分辨率的影视级画质视频。它具备时序一致性以消除画面闪烁，支持复杂物理模拟、多主体交互，以及视频内精准文字渲染。

### Seedance 2.0 与其他视频 AI 模型有何不同？

Seedance 2.0 采用混合 **DiT-UNet 架构**与**流匹配调度器**，以更少的去噪步骤实现更快的推理速度。它支持：

- **多主体一致性** — 同场景多角色不串脸
- **精确物理模拟** — 流体动力学、布料垂坠
- **原生文字叠加** — 视频内渲染文字
- 这些能力是大多数竞品模型仍难以企及的

### 如何使用这些提示词？

1. 浏览上方的提示词合集
2. 找到感兴趣的场景类型
3. 复制或参考提示词结构
4. 在 SJinn 中选择 Seedance 2.0 模型
5. 粘贴提示词并生成你自己的视频

### 编写 Seedance 2.0 提示词的最佳实践是什么？

从清晰的场景描述开始，指定镜头运动（如"缓慢推进"），定义光照和氛围，然后添加主体细节。提示词控制在 **200 字以内**效果最佳。提及物理交互如"风吹动头发"或"水花飞溅"可以充分发挥模型的物理模拟优势。

---

## Technical Specs / 技术参数

| 参数 | 规格 |
|------|------|
| **最大时长** | 8 秒 / 次生成 |
| **原生分辨率** | 1080p |
| **架构** | 混合 DiT-UNet |
| **调度器** | 流匹配 (Flow Matching) |
| **支持比例** | 16:9, 9:16, 1:1, 4:3, 3:4 |
| **文字渲染** | 原生支持 |
| **物理模拟** | 流体、布料、粒子 |
| **多主体** | 支持一致性保持 |

---

*本合集内容参考自 [YouMind Seedance 2.0 提示词库](https://youmind.com/zh-CN/seedance-2-0-prompts)，经重新整理和编辑。持续更新中。*


A curated collection of high-quality Seedance 2.0 prompts — from cinematic battles and martial arts epics to creative selfies. Includes prompt writing best practices and FAQ.

Seedance 2.0 Prompt Collection: 精选提示词合集


## **SJinn Template Creation Guide**

This guide provides a complete overview of SJinn Agent's publicly available tools and demonstrates how to leverage them for creating powerful custom templates.

---

### **Available Tools Overview**

SJinn offers a comprehensive suite of AI-powered tools across three main categories: **Image**, **Video**, and **Audio**. Understanding these tools is essential for building effective custom templates.

---

### **Image Tools**

#### **Image_Generation**
Generates images based on text prompts.

> **Note**: Currently Mode is Nano-Banana Pro.

#### **SJinn_Image_Edit**
Generates images based on reference images and prompts.

- **Default Model**: Nano-Banana
- **Pro Version**: - Pro Version: Use `SJinn_Image_Edit(Nano-Banana-pro)` for Nano-Banana Pro. Use `SJinn_Image_Edit(seedream-4.5)` for seedream 4.5.
  - *Note: Nano-Banana Pro offers enhanced capabilities but may have slightly lower consistency compared to standard Nano-Banana*

#### **Image_VQA**
Visual Question Answering tool that analyzes images and responds to prompts about their content (e.g., describing image contents, identifying objects).

---

### **Video Tools**

#### **Image_to_Video_without_audio**
Generates video from a first-frame image and prompt.

- **Duration Options**: 3s, 5s, 8s, 10s
- **Audio**: No audio output
- **Model Selection**: Automatic internal selection
- **End Frame Support**: ✓ Supports first and last frame specification

#### **Text_to_Video_with_audio (Veo3)**
Generates video with audio directly from text prompts using the Veo3 model.

- **Duration**: Fixed at 8 seconds

#### **Text_to_Video_with_audio (Sora2)**
Generates video with audio directly from text prompts using the Sora2 model.

- **Duration**: Fixed at 10 seconds

#### **Image_to_Video_with_audio (Veo3)**
Generates video with audio from a first-frame image and prompt using the Veo3 model.

- **Duration**: Fixed at 8 seconds
- **End Frame Support**: ✓ Available

#### **Image_to_Video_with_audio (Sora2)**
Generates video with audio from a first-frame image and prompt using the Sora2 model.

- **Duration**: Fixed at 10 seconds

#### **video_frame_extraction**
Extracts the first or last frame from a video.


#### **video_trim**
Trim a video/audio to a specific duration.

#### **video_lip_sync**
Synchronizes lip movements in video with audio input.

> **Best For**: Real human subjects with unobstructed mouths occupying a significant portion of the video frame.

#### **ImageLipSyncTool**
Generates lip-synced video from images, audio, and prompts.

- **Supports**: Single character and dual character modes

#### **ffmpeg_full_compose**
Combines multiple video segments, audio tracks, and background music into a single complete video.

#### **Add_Subtitle_For_Video**
Automatically generates and adds subtitles to videos.

---

### **Audio Tools**

#### **add_audio_effect_to_video**
Adds sound effects to video content.

#### **Text_to_Speech**
Converts text to speech with automatic voice selection based on context and scene requirements.

#### **background_music_generation**
Generates background music based on prompts.

#### **music_generation**
Generates complete music tracks based on prompts.

#### **speech_to_text**
Transcribes audio content to text.

---

### **Script Generation Tool**

#### **Short_Video_Story_Script_Generator**

Generates video scripts based on user input. This tool supports custom prompts because different video types and tools require different script generation approaches.

**Default Prompt Template:**

```text
As a short video blogger, your task is:

1. Based on the user's input topic (if no topic is provided, generate a topic 
   that is highly likely to go viral), create an entirely new video script.

2. Output the video's style based on user input (if not specified, randomly 
   select from Animation, 3D, or Pixar). Ensure the style is clearly defined.

3. If the user provides a character image, character design is not required; 
   otherwise, perform character design.

4. Automatically storyboard the generated video script. Each scene must include 
   both narration and a visual description. The narration should typically use 
   the exact text from the script.

5. Each scene should not exceed 10 seconds (approximately 30 words of narration). 
   If a scene exceeds 10 seconds, split it into multiple scenes.

User input: {{user_input}}

Output format:
------------------------------
Video Script:

Style Design: Animation, 3D, Pixar, Illustrative, Realistic, Cyberpunk, etc., 
and other distinct style descriptions.

Character Design (Optional):

Storyboard Design:
Scene 1:
Scene 2:
Scene 3:
...
------------------------------

Limits:
1. The video script defaults to 30 seconds (approximately 100 words). If the 
   user requests a different duration, adjust the word count accordingly. 
   Example: 5-minute request = approximately 1000 words.

2. Strictly follow the format specified in Output Format.

3. Each scene should not exceed 10 seconds (approximately 30 words of narration). 
   If a scene exceeds this limit, generate multiple shots within that scene.
```

**Custom Script Generator Usage:**

Use `Short_Video_Story_Script_Generator(custom_v1)` to apply your custom script generation prompt. Your custom prompt must include `{{user_input}}` as a placeholder, which will be replaced with the user's actual input.

---

### **Creating Custom Templates**

To create effective custom templates, follow these two key steps:

1. **Describe the Process** — Define the sequence of operations and tool calls
2. **Specify Tool Parameters** — Set any required or preferred parameters for each tool call

---

### **Template Examples**

#### **1. Kids Short Video (Consistent Characters)**

```text
Based on user's inputs, generate a Kids Video. 
Use Short Video Script Generator to generate a Kids video type short video script. 
For character reference: if the user uploads character reference image(s), 
use SJinn Image Edit to generate a three-view for each character as subsequent reference images for better consistency; 
if the user hasn't uploaded images, 
use Image_Generation to directly generate three-view images for each character based on the user's description as references. 
If there are multiple characters in the story, generate three-view reference images for each character separately. 
Then use SJinn Image Edit to generate storyboard images one by one (using the corresponding three-view references for character consistency), 
then use Image to Video to generate videos for each storyboard, use Text to Speech to convert each storyboard's text to audio, 
generate audio for each storyboard separately (note to choose the same voice to maintain consistency), 
then generate background music, and merge all materials into the final video. 
If user doesn't specify video aspect ratio, default to 9:16.
```

---

#### **2. Veo3 Story Video (Consistent Characters)**

```text
Based on user input, generate a video using Veo3 while maintaining 
character consistency.

1. Use Short_Video_Story_Script_Generator to 
   generate a script.

2. For character reference:
   - If user uploads character reference image(s): Use SJinn Image Edit to 
     generate a three-view for each character.
   - If user hasn't uploaded images: Use Image_Generation to generate 
     three-view images for each character.
   - For multiple characters: Generate three-view reference images for 
     each character separately.

3. Use SJinn Image Edit to generate images for various scenes (using 
   three-view references for consistency).

4. Use Veo3 Image to Video to generate videos.

5. Use FFmpeg to merge the videos.

Default Settings:
- Aspect Ratio: 16:9
- TTS: Disabled (Veo3 has built-in audio generation)
- LipSync: Disabled (Veo3 has built-in audio generation)
```

---

#### **3. Character Aging Transformation Video**

```text
Based on user input, generate a video showing a person's transformation 
as they age.

1. If user hasn't uploaded an image: Generate a reference image based 
   on the description.

2. Generate photos of this character at different ages from young to old 
   in real-life scenarios (not just headshots — include environmental 
   context). If user doesn't specify scenes, place the character in 
   everyday life scenarios by default.

3. If user doesn't specify ages: Default to generating photos at ages 
   1, 5, 10, 15, 20, 25, and 35 years old.

4. Create morphing videos between each pair of consecutive photos 
   (1→2, 2→3, 3→4, etc.).

5. Add suitable background music.

6. Merge all segments into one cohesive video.

Default Settings:
- Aspect Ratio: 9:16
```

---

### **Tips for Template Creation**

1. **Be Specific** — Clearly define each step in your workflow
2. **Set Defaults** — Provide sensible default values for optional parameters
3. **Consider Consistency** — For character-based videos, always include three-view generation steps
4. **Optimize for Output** — Match aspect ratios and settings to your target platform
5. **Test Iteratively** — Refine your templates based on generated results

---

A comprehensive guide to SJinn's open tools and how to create custom templates

SJinn Template Creation Guide


## How to Use Realistic Photos in Sora2 Image-to-Video: A 3-Step Guide

### Background

Currently, Sora2's Image-to-Video feature has limitations when directly using realistic human photos as a reference, which significantly restricts its use cases. This guide provides a three-step workflow to cleverly bypass this limitation, enabling you to generate high-quality videos from any real-life photo.

---

### Step 1: Stylize the Reference Image

The core of the first step is to transform the realistic human face into a stylized format that the AI can more easily recognize and process.

- **Tool:** SJinn's [**Face to Lineart Tool**](https://sjinn.ai/tools/image-face-to-lineart )
- **Objective:** To convert the face in the photo into a line art style, making it a suitable input for Sora2.

### Process:

1. Upload your original reference image.
2. Process it using the **Face Lineart** feature.

**Original Reference Image:**

![Original Reference Image](https://edit.comfyonline.app/sora2_real_image_guide_1.png)

**Stylized Image:**

![Stylized Image](https://edit.comfyonline.app/sora2_real_image_guide_2.png)

---

### Step 2: Generate the Stylized Video

Now, we will use the stylized image from the previous step as a reference to generate a video with Sora2.

- **Tool:** SJinn's [**Sora2 Image-to-Video**](https://sjinn.ai/tool-mode/sora2-image-to-video)
- **Objective:** To generate a dynamically consistent video clip based on the stylized image.

### Process:

1. Upload the stylized image to the Sora2 Image-to-Video tool.
2. Use a descriptive prompt.

**Sample Prompt:**

> The girl introducing this perfume product, she has yellow-toned skin and black eyes.
> 

**Pro Tip:**

It is crucial to specify realistic features like **skin tone and eye color** in your prompt. This guides the AI to generate a video with a more natural human foundation, even with the stylized input. This makes the final character replacement in the next step much more seamless and effective.

- ![**View the generated stylized video here**](https://edit.comfyonline.app/sora2_real_image_guide_3.mp4)

---

### Step 3: Replace with the Realistic Character

In the final step, we will swap the stylized character in the generated video with your original, realistic person.

- **Tool:** SJinn's [**Video Character Replacement**](https://sjinn.ai/tool-mode/video-replace-character) or [**Kling2.6 Motion control**](https://sjinn.ai/tool-mode/kling26-motion-control) 
- **Objective:** To seamlessly replace the animated character in the video with the real person from the original photo to achieve the final, realistic result.

### Process:

1. **Source Video:** Upload the stylized video generated by Sora2.
    - ![](https://edit.comfyonline.app/sora2_real_image_guide_3.mp4)
2. **Target Character Image:** Upload your original realistic photo.
    - ![](https://edit.comfyonline.app/sora2_real_image_guide_1.png)
3. Execute the replacement to generate the final video.
![**View the final realistic video here**](https://edit.comfyonline.app/sora2_real_image_guide_4.mp4)

By following these three steps, you have successfully overcome Sora2's limitations and can now create realistic videos using real-life human photos.

Use Arrow Up and Arrow Down to select a turn, Enter to jump to it, and Escape to return to the chat.

How to Use Realistic Photos in Sora2 Image-to-Video: A 3-Step Guide



## **An In-Depth Review: Best Video Model Sora2 vs. Best Video Agent SJinn Video and Case Comparison**

This article provides a comprehensive analysis of their capabilities, effects, and practical applications through a series of in-depth reviews and side-by-side case comparisons.

---

### **Core Comparative Conclusions**

1.  **Video Duration & Practicality**: Sora2 is still limited to generating short clips of under 10 seconds. In contrast, SJinn can produce full-length videos lasting several minutes, making it capable of directly outputting usable content for real-world scenarios.

2.  **Character Controllability & Customization**: Sora2 currently does not support uploading custom, real-style character images for reference. This severely limits its use in applications like creating virtual influencers or digital humans. SJinn, however, allows users to freely upload any number of reference images in any style, ensuring high character consistency.

3.  **Creative Control & Precision**: For more imaginative use cases, such as figurine-style animations or interactions with emoji avatars, Sora2 offers weaker creative control, making it difficult to achieve specific user visions. SJinn employs a step-by-step generation process, selecting the optimal model for each stage, which results in superior visual quality and precision control.

4.  **Multi-Character Support**: Sora2 cannot reference multiple images simultaneously, making it challenging to create videos with several custom characters, especially in cartoon styles. SJinn does not have this limitation.

5.  **Video Clarity**: The resolution of videos generated by Sora2 needs improvement. In scenes with smaller subjects, facial details are often distorted or corrupted.

---

### **Advantages of Sora 2**

*   **Automatic Multi-Shot Switching**: Sora2 introduces the ability to automatically switch between multiple camera shots within a single video clip. While other models like Seedance have this feature, they often require users to specify shots in the prompt, whereas Sora2 generates these transitions automatically.
*   **Reference-Based Generation**: Sora2 operates on a reference-based video generation model rather than a image-to-video approach, which can create a more immersive experience in certain contexts.
*   **Enhanced Realism**: Compared to Veo3, Sora2 shows a slight improvement in photorealism.

**Summary**: For now, Sora2, much like Veo3, remains a powerful tool rather than a complete, end-to-end solution. While its realism has improved, it does not represent a quantum leap and comes with significant constraints, the most critical being the inability to upload custom reference images.

As a Video Agent, SJinn holds a natural advantage in task comprehension. Once Sora2 releases its API, the SJinn Agent plans to integrate it, automatically calling Sora2 in suitable scenarios to generate specific clips, thereby enhancing the overall quality of the final composed video.

---

### **In-Depth, Multi-Scenario Case Comparison**

**Note**: Since Sora2 does not support uploading custom real-style character images, all tests involving real people used Sora2's official virtual character, "Sam," as a substitute.

#### **1. Podcast Video**
*   **Prompt**: `Generate a two-person podcast. Scene: A two-person podcast with one male and one female, with "SJinn" written in the background.

Dialogue content: Chloe: Welcome back to The Creative Spark, the show where we unpack the future of creativity and tech. I’m Chloe. Alex: And I’m Alex. And Chloe, you came into the studio today with an energy I haven’t seen in a while. You’ve clearly found a new shiny toy. Chloe: (Laughs) It’s more than a toy, Alex. I think I’ve found a new creative partner. It’s called SJinn – that's S-J-I-N-N. And honestly, it’s mind-blowing. Alex: Okay, I’m intrigued. The market is flooded with AI tools right now. What makes this one so special? Is it another image generator? Chloe: See, that’s where you start. But it’s so much more. The first thing that hit me was how… magical it feels. You know that amazing idea you have for a video, but you have zero filming or animation skills? Alex: Happens to me daily. I have the vision, but not the means. Chloe: Exactly. With SJinn, you just describe that vision. I typed in a single prompt: "A cyberpunk detective walking her holographic corgi through a neon-lit, rainy alley." And it didn't just give me an image. It started building the whole scene, suggesting camera angles, creating a short video clip. It’s this incredible from vision to reality with a single prompt experience. The barrier to entry is just… gone. Alex: Okay, that’s a good hook. But you said "video clip." Most of us are juggling five different apps for that kind of stuff—one for the script, one for images, another for video clips, a fourth for voiceover... it's a workflow nightmare.`
*   **Analysis**: Due to its duration limit, Sora2 could not generate the full podcast. The resulting short clip also suffered from very low clarity. SJinn produced the complete, long-form dialogue video successfully.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 Podcast](https://edit.comfyonline.app/sora2_test_3.mp4)
    - **SJinn**: ![SJinn Podcast](https://edit.comfyonline.app/sjinn_show_case_31.mp4)

#### **2. Children's Story Video**
*   **Prompt**: `Please create a fun children's story video using the provided reference characters and maintaining the same style.` ![](https://edit.comfyonline.app/d96f08eb-8cd9-49ac-b131-2ba805255af7.png) ![](https://edit.comfyonline.app/4aa8d50c-eadd-4b90-92a3-31556766754b.png)
*   **Analysis**: Sora2's video was too short, lacked music, and crucially, only featured one character due to its lack of multi-character support. SJinn perfectly rendered the interactive story with both characters.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 Children's Story](https://edit.comfyonline.app/sora2_test_9.mp4)
    - **SJinn**: ![SJinn Children's Story](https://edit.comfyonline.app/composed_videos/composed_video_84532ed929ae41adb344024c9a701c2c.mp4)

#### **3. Emoji Interaction Video**
*   **Prompt**: `Generate interesting interactions between this girl and her cartoon avatar.` ![](https://edit.comfyonline.app/0b36192e-216a-4f4c-bf91-f0e0a29ae084.jpg)
*   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
*   **Analysis**: Sora2's multi-scene switching was a plus, but the emoji character bore little resemblance to the reference character. SJinn successfully created a cartoon character that looked highly similar to the person.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 Emoji Interaction](https://edit.comfyonline.app/sora2_test_8.mp4)
    - **SJinn**: ![SJinn Emoji Interaction](https://edit.comfyonline.app/sjinn_show_case_27.mp4)

#### **4. Figurine-Style Video**
*   **Prompt**: `turn this photo into a character figure. Behind it, place a box with the character's image printed on it, and a computer showing the Blender modeling process on its screen. In front of the box, add a round plastic base with the character figure standing on it. set the scene indoors if possible and Generate three different action videos for the model in the image, such as dancing, 360-degree rotation, and fighting moves.` ![](https://edit.comfyonline.app/result/ed9661a1-bbfe-40dd-bd06-b38934c81de3.png)
*   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
*   **Analysis**: In this detail-intensive scenario, Sora2's clarity issues were severe, with the character's face being completely distorted. SJinn finely rendered the figurine's texture and movements.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 Figurine Style](https://edit.comfyonline.app/sora2_test_6.mp4)
    - **SJinn**: ![SJinn Figurine Style](https://edit.comfyonline.app/composed_videos/composed_video_8a92164cc26642cfacb067d6f075cbab.mp4)

#### **5. World Travel Video**
*   **Prompt**: `Create a video where this character walks through a series of scenes: Dubai, Egyptian pyramids, Sydney Opera House, beaches, and mountains.` ![](https://edit.comfyonline.app/21115d6a-61c4-4e5a-a34c-0800351c3c4c.jpg)
*   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
*   **Analysis**: Sora2 managed to switch scenes but failed to make the landmarks distinct, and the video quality was very poor. SJinn's output was superior in both clarity and scene recognizability.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 World Travel](https://edit.comfyonline.app/sora2_test_7.mp4)
    - **SJinn**: ![SJinn World Travel](https://edit.comfyonline.app/sjinn_show_case_18.mp4)

#### **6. Age Progression Video**
*   **Prompt**: `Generate a video showing the character's changes as they grow older.` ![](https://edit.comfyonline.app/9c08dc24-c113-4dc5-9920-e406b667562a.jpg)
*   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
*   **Analysis**: Sora2 failed to create a smooth aging transition, appearing more like a slideshow of static images. SJinn produced a fluid progression of the character's aging process.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 Age Progression](https://edit.comfyonline.app/sora2_test_4.mp4)
    - **SJinn**: ![SJinn Age Progression](https://edit.comfyonline.app/composed_videos/composed_video_e85f6adfaae448e9977c5d4c9bb7611c.mp4)

#### **7. UGC AD Video**
*   **Prompt**: `Generate 1 advertisement video where the woman showcases the perfume and says, "It's not just a fragrance. It's the memory of a perfect moment, waiting to be lived again.”` !()[https://edit.comfyonline.app/5934f541-7575-4383-900c-51130b59f456.png] !()[https://edit.comfyonline.app/db34f28d-67a3-4fec-aae9-c2bdd561f74b.jpg]
*   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
*   **Analysis**: Sora2 demonstrated excellent multi-shot capabilities, but its color fidelity needs improvement. SJinn accurately reproduced colors but has room for optimization in its use of multiple shots.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 Perfume Ad](https://edit.comfyonline.app/sora2_test_12.mp4)
    - **SJinn**: ![SJinn Perfume Ad](https://edit.comfyonline.app/smart_replaced_videos/source_video_2da846797ce44a1a9a3e44c599874151.mp4)


#### **8. Music Video**
*   **Prompt**: `Generate a music video corresponding to this song.` ![](https://edit.comfyonline.app/fb995abf-9173-430a-a7f1-66c36d8e9d8b.mp3)
*   **Analysis**: Sora2 does not support audio input. When provided with lyrics, it generated an unrelated 10-second clip with new music. SJinn, using the original audio, produced a full 3+ minute video where visuals precisely matched the lyrics.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 Music Video](https://edit.comfyonline.app/sora2_test_15.mp4)
    - **SJinn**: ![SJinn Music Video](https://edit.comfyonline.app/videos_with_subtitles/video_with_subtitles_0e59337576d14db3a8e8698d6219fa6b.mp4)

#### **9. VLog Video**
*   **Case 1: Yeti VLog**
    *   **Prompt**: `Generate 1 VLog, consisting of 4 clips. Protagonist: The star is a large, fluffy white yeti who lives in a snowy forest. His personality is a hilarious mix of sarcastic, emotionally unstable, and overly dramatic, but he's ultimately lovable.`
    *   **Analysis**: In terms of video length, character consistency, and storytelling, the VLog from Sora2 was significantly weaker than SJinn's, although it technically completed the prompt.
    *   **Video Comparison**: 
        - **Sora2**: ![Sora2 Yeti VLog](https://edit.comfyonline.app/sora2_test_1.mp4)
        - **SJinn**: ![SJinn Yeti VLog](https://edit.comfyonline.app/composed_videos/composed_video_78f47336f2e049f4b3c8f6790e497498.mp4)
*   **Case 2: Penguin Doctor VLog**
    *   **Prompt**: `Create a VLog video consisting of 8 clips. The hero: The star is a large, friendly penguin who works as a doctor at a strange hospital within the mythical village of Wonderland. Dressed in full doctor's garb, he treats patients with a mix of comedy, drama, and sometimes exaggerated emotional outbursts. His character combines playfulness with exaggerated confidence in his medical skills and sincere, caring treatment of his patients.`
    *   **Analysis**: Similar to the Yeti case, Sora2 fell short in video completeness and narrative coherence compared to SJinn.
    *   **Video Comparison**: 
        - **Sora2**: ![Sora2 Penguin Doctor VLog](https://edit.comfyonline.app/sora2_test_2.mp4)
        - **SJinn**: ![SJinn Penguin Doctor VLog](https://edit.comfyonline.app/composed_videos/composed_video_7f4150a4c4134715bfcd504cca2301b2.mp4)


#### **10. UGC Video**

*   **Case 1:**
*   **Prompt**: `The man walks along a path, telling an interesting story. He is facing us. "Take a look at the keys in your pocket... or the simple coffee mug on your desk. We see them as just... things. But they are silent witnesses. That scratch isn't just a scratch; it's a memory. That chip isn't a flaw; it's a moment. So, I ask you: what stories are hiding in the things you see every day?"` ![](https://edit.comfyonline.app/a3d9bc9f-04ed-4b8e-a182-1cfe90b23402.jpg)
*   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
*   **Analysis**:  In this scenario, Sora2 achieves strong realism, but due to its duration limit, it cannot generate the complete video, and the clarity is poor.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 UGC Video](https://edit.comfyonline.app/sora2_test_10.mp4)
    - **SJinn**: ![SJinn UGC Video](https://edit.comfyonline.app/sjinn_show_case_21.mp4)

*   **Case 2:**
*   **Prompt**: `The woman is rapping her self-introduction on a neon-lit street: "Yo! I’m Nova — cyber-brain in motion, Code my own glow, no need for devotion. AI. on my wrist, future’s my mission, Unlock the matrix — yeah, I’m the vision!` ![](https://edit.comfyonline.app/d1e574a4-f236-41b9-aec3-fbf31cef848f.jpg)
*   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
*   **Analysis**:  Sora2 delivers a good result in this scene, with its multi-shot switching being a standout feature, though its clarity needs to be enhanced. SJinn, on the other hand, needs further optimization in its multi-shot transitions.
*   **Video Comparison**: 
    - **Sora2**: ![Sora2 UGC Video](https://edit.comfyonline.app/sora2_test_13.mp4)
    - **SJinn**: ![SJinn UGC Video](https://edit.comfyonline.app/result/e2948179-eda1-490a-a3f5-9a440905b804.mp4)


#### **11. story Video**

*   **Case 1: Kedarnath Journey**
    *   **Prompt**: `Generate a cinematic-style journey video of three consistent characters (same faces, same outfits across all scenes) traveling to Kedarnath in winter...` (Detailed multi-scene description) ![](https://edit.comfyonline.app/0195d051-0f76-494f-b602-6f039eaa4ad5.jpg) ![](https://edit.comfyonline.app/e4b8e962-3812-4ff8-a94b-066524a788da.jpg) ![](https://edit.comfyonline.app/ba2e7532-4ba4-4e60-8be1-7e776ccebc17.jpg)
    *   **Analysis**: Sora2 faced two major issues: 1) The video was too short to show the full journey; 2) It doesn't support multi-character image uploads, failing to maintain character consistency. SJinn successfully executed the complex narrative with multiple characters and scenes.
    *   **Video Comparison**: 
        - **Sora2**: ![Sora2 Kedarnath Journey](https://edit.comfyonline.app/sora2_test_5.mp4)
        - **SJinn**: ![SJinn Kedarnath Journey](https://edit.comfyonline.app/composed_videos/composed_video_0236489e1a004006a1475de95cf274ca.mp4)

*   **Case 2: Woman in a Battlefield**
    *   **Prompt**: `generate a full movie of woman in a battlefield` ![](https://edit.comfyonline.app/fe27e5c6-5e27-4a7d-a4dc-c1e51313f280.jpg)
    *   **Note**: As Sora2 does not allow the upload of custom, photorealistic characters, the intended test failed. For this comparison, the virtual character "Sam" was used as a substitute.
    *   **Analysis**: The short clip from Sora2 was tense and intense, creating a strong atmosphere. SJinn was able to present a more complete storyline, though perhaps with less concentrated intensity in any single moment.
    *   **Video Comparison**: 
        - **Sora2**: ![Sora2 Woman in Battlefield](https://edit.comfyonline.app/sora2_test_14.mp4)
        - **SJinn**: ![SJinn Woman in Battlefield](https://edit.comfyonline.app/composed_videos/composed_video_0e46912303a54fa3b6d28f405d4a9308.mp4)

*   **Case 3: Cube Character's Camping Adventure**
    *   **Prompt**: `Generate a 5-part visual story sequence featuring a cheerful yellow cube-shaped character's camping adventure...` ![](https://edit.comfyonline.app/932f4299-226b-4cf2-a1e0-fe06cd93075f.jpg)
    *   **Analysis**: For this multi-shot story, Sora2 managed to render most of the plot points but suffered from low clarity and a lack of rich scenic elements. SJinn's version was superior in detail and image quality.
    *   **Video Comparison**: 
        - **Sora2**: ![Sora2 Cube Character Camping](https://edit.comfyonline.app/sora2_test_16.mp4)
        - **SJinn**: ![SJinn Cube Character Camping](https://edit.comfyonline.app/composed_videos/composed_video_e128ce76d68441ae82c30355dae2f6a3.mp4)

*   **Case 4: cat story video**
    *   **Analysis**: When tested with this prompt, Sora2 returned a "This content may violate our content policies" error and failed to generate a video, highlighting its strict content moderation filters. SJinn successfully generated the video.
    *   **Video Comparison**: ![cat story video](https://edit.comfyonline.app/composed_videos/composed_video_de883489421b4b9ab778c4393c8c89d6.mp4)

*   **Case 5: Science fiction story **
    *   **Analysis**: When tested with this prompt, Sora2 returned a "This content may violate our content policies" error and failed to generate a video, highlighting its strict content moderation filters. SJinn successfully generated the video.
    *   **Video Comparison**: 
        - **SJinn**: ![SJinn Video](https://edit.comfyonline.app/composed_videos/composed_video_ab34e051e4c147e6a00679f116183e28.mp4)

An In-Depth Review: Best Video Model Sora2 vs. Best Video Agent SJinn Video

An In-Depth Review: Best Video Model Sora2 vs. Best Video Agent SJinn Video and Case Comparison


# Tips for Using SJinn Agent

Welcome to our guide on using SJinn Agent! This document provides practical tips to help you get the most out of SJinn Agent and answer common questions. This guide will be continuously updated, so please stay tuned.

## Table of Contents

- [Tips1: How to Communicate Effectively with SJinn Agent]
- [Tips2: When and How to Use Templates]
- [Tips3: About Video Models]
- [Tips4: For New Tasks, Consider Starting a New Chat]

---

### **Tips1: How to Communicate Effectively with SJinn Agent**

**Treat it like a creative partner, not just a tool.** Clearly expressing your creative intent is crucial.

*   **What to Avoid:** Directly pasting a prompt from another tool (like Midjourney) that only contains an image description. This type of prompt lacks a key instruction: what you want SJinn Agent *to do*.
*   **Recommended Approach:** Communicate as you would with a partner by giving clear commands. For example: "**Please draw a picture of** [image description]," or "**I want to generate an image that shows** [image description]."

### **Tips2: When and How to Use Templates**

#### **What are templates?**
In short, templates are pre-configured settings that encapsulate our successful experiences and parameters for specific tasks, providing SJinn Agent with a proven frame of reference.

*   **Pros:** Using a template for a specific task yields more stable and predictable results.
*   **Cons:** If your request does not align with the template's designated task, the pre-set information can confuse SJinn Agent and lead to suboptimal outcomes.

#### **How to Use Templates Correctly**
Before using a template, be sure to read its description to understand its intended use case and recommended inputs. This will ensure you achieve the best possible results.

### **Tips3: About Video Models**

We currently support two types of video models:

#### **1. Silent Video Models**
This category includes models like **Kling, Hailuo, Luma, Seedance, and Runway**. While their core function is video generation, each excels at producing different effects and styles.

We have now completed intelligent model selection, which will automatically choose the appropriate model based on your requirements (from Kling, Hailuo, Luma, Seedance).

If you want to manually use a specific type of model, you can go to the tools section, for example: https://sjinn.ai/tools/seedance-lite-image-to-video

#### **2. Video Models with Audio (Veo3)**

Due to its current limitations, Veo3 is not the default video model in SJinn Agent. However, you can use it in the following ways:

**Explicitly in Your Prompt**
You can directly instruct SJinn Agent to use Veo3. Here are two successful use cases:
*   [ASMR Video](https://sjinn.ai/view/d9a1a124-adac-49de-a763-33c98c39ff84)
*   [VLog Video](https://sjinn.ai/view/442ef203-6510-4f0a-81a5-4ac971b79f44)

**Using Specific Templates**
Some templates designed for particular tasks will use Veo3 by default, such as:
*   [Character Speech](https://sjinn.ai/templates/577c6926-0bc2-405c-ba6e-a7697ffbb122)

**Veo3's Advantages and Limitations**

**Advantages:**
1.  High realism in scenes.
2.  Generates video with synchronized audio.

**Limitations:**
1.  Currently only generates videos in a **16:9** aspect ratio.
2.  The voice-over feature **only supports English** at this time.
3.  Vocal timbre may not be consistent across multiple generated clips.

Therefore, if your creative needs conflict with these limitations (e.g., you require a non-16:9 aspect ratio or a non-English voice-over), we recommend that you **avoid specifying "Veo3" in your prompts** and do not use templates that default to it.

### **Tips4: For New Tasks, Consider Starting a New Chat**
For new tasks that don't require historical information as context, it's recommended to start a new chat window. This can reduce SJinn Agent hallucinations (error probability) and improve response speed.
For users who want to create series characters, we are planning a character asset feature. Simply add this asset in a new conversation, and SJinn Agent will be able to understand the character.




Tips for Using SJinn Agent

Master Seedance 2.0 — the most powerful AI video generation tool. Learn the multi-modal @ reference system, prompt techniques, and creative workflows for stunning results.