Tomato AI LogoTomato AI
Home
Video AI
Pricing-50%
Blog⚽ World CupHOT
←
Tomato AI LogoTomato AI

Tomato AI integrates Jimeng 3.0, Veo 3.1, Sora 2, Kling 3 and other top models. Deliver commercial-grade videos from text, images or video in seconds.

Product

  • Text to Video
  • Image to Video
  • 关于我们

Resources

  • Pricing
  • FAQ
  • Blog

© 2026 • Tomato AI All Rights Reservedsupport@tomato.ai
Terms of ServicePrivacy Policy
Tomato AI is an independent product and is not affiliated with ByteDance, Google, OpenAI, etc.
← Back to Blog
Seedance 2.0

Standards: Language, Special Characters & Advanced Formula Guide

2026-06-098 min readTomato AI Team

Standards: Language, Special Characters & Advanced Formula Guide


1. Language Standards

Core Rules

Dialog language must be consistent — avoid mixing Chinese and English (except for proper nouns).

#### Correct Examples

{The weather is nice today}  ✓
{How are you}   ✓  (English can be used in English-language scenes)

#### Incorrect Examples

{The weather is nice today, you know}  ✗  (Mixing Chinese and English)

Non-English Dialog Scenarios

If the dialog is in a non-English language (neither Chinese nor English), specify the language.

#### Examples

Said in Japanese: {こんにちは}
Said in Korean: {안녕하세요}

2. Special Characters Standards

Overview

Using symbols appropriately in prompts helps the model accurately understand different types of information.

Symbol Reference Table

Information TypeSymbolExample
Music()(Upbeat rock music playing in the background)
Sound Effect<>
Dialog{}{Hello, world}
Subtitle【】【Chapter 1: Departure】

Usage Examples

<Crisp metallic scraping sound>           # Sound effect
(Background music crescendo)              # Music
{How did the exam go, did you pass}        # Dialog
【Chapter 1: Departure】             # Subtitle

Non-English Dialog Rules

If the dialog is in a non-English language (neither Chinese nor English), specify the language:

Said in Japanese: {こんにちは}

3. Advanced Prompt Formula Guide

Core Concept

Seedance 2.0 is essentially a multimodal AI director: it reads your text prompt, images, video, and audio simultaneously, and internally decomposes them into the "spatial layer" (what's in the frame) and the "temporal layer" (how things change over time) to understand and generate visuals.

Therefore, a good prompt is not just "copywriting-style description" but rather "engineering-style instructions."

Advanced Formula

!Advanced Formula Diagram

Advanced Prompt Formula Structure Diagram

!Prompt Template Reference

Prompt Element Template Reference

Advanced Prompt Formula:

Precise Subject + Action Details + Scene Environment + Lighting & Color +
Camera Movement + Visual Style + Quality + Constraints

!Common Issues Troubleshooting Flow

Common Issues Troubleshooting & Optimization Flow

Writing Process

  • Identify "who" is doing "what" → Define subject + action description
  • Set "where" and "what atmosphere" → Scene environment + lighting & color
  • Tell the model "how to shoot" → Camera movement
  • Refine the result → Style + quality + constraints

Summary in One Sentence

First identify who is doing what, then set where and what atmosphere, then tell the model how to shoot, and finally refine the result with style, quality, and constraints.

🍅 Try AI Video Generation Free on Tomato AI

Sign up for free credits. Access Seeddance 2.0, Sora 2, Kling 3 & more top models. No watermark, 1080P output.

Start Creating Free →