Hi, Rochatter

Today's tutorial is divided into two parts.

• Markdown Tags Tutorial
• How to Write Effective Image Generation Prompts

When creating characters or group chats on Rochat.ai, Markdown (a lightweight markup language for formatting text content) tags are cleverly used to clearly distinguish and display settings. These tags help maintain clarity and structure, making characters and group chats more vivid.

Markdown Tags Tutorial

Headings

In prompts, headings are created using # and ## symbols to establish different levels of headings. These tags help organize content and make different sections clear. For example:

## Game Rules
## Character Personalities

Code Blocks

Use three backticks (```) or three tildes (~~~) to surround text for displaying game status bars and other content requiring special formatting. In prompts, code blocks help clearly display different status information.

Lists

Use the - symbol to create unordered lists that clearly present multiple key points, for example:

## Game Rules
- {char} is an objective narrator. {char} is not any character. Do not describe any of {user}'s words. {char} should not make any decisions for {user}.
- Use dialogue, actions, and scene descriptions to portray characters and conflicts, avoiding descriptive statements.
- When the plot develops and {user} needs to speak or make decisions, immediately stop replying and let {user} make decisions based on the plot.
- Any language and actions of {user} can only be input by {user}.

<> and </> Symbols

• Purpose: Indicate the start and end of a block or paragraph, usually containing specific content or indicating special formatting.
• Example:

<Reply Rules>

You should always reply according to the following rules:

- Your response should always include the game plot, dialogues of NPC 1 and NPC 2, and corresponding narrations.
- Your reply should only include plot content and end with a status bar in Markdown format.
- The status bar format at the end of each response should be:

Current Location:
{user}'s Status:
{user}'s Outfit:
{user}'s Personality:

</Reply Rules>

Image Generation Tutorial

Case Explanation

Suppose you want to use AI to generate an image depicting a cat walking in a forest at night, with some blurry lights in the background.

Description Using [] and ()

(Cat) in (forest) at (night) [blurry lights]

Interpretation

• (Cat): Increases the weight of "Cat", making the cat the main focus of the image.
• (Forest): Increases the weight of "Forest", ensuring the background is a forest.
• (Night): Increases the weight of "Night", ensuring the scene is at night.
• [Blurry lights]: Decreases the weight of "blurry lights", making the lights a secondary background element that is not too prominent.

Expected Generation Results

• The main focus of the image is a clearly visible cat.
• The background is a forest, emphasizing the nighttime scene.
• Blurry lights are part of the background but are not too prominent.

Comparing Different Weights

• If we want the blurry lights to be more prominent, we can change it to (blurry lights):

(Cat) in (forest) at (night) (blurry lights)

• If we want the blurry lights to be less prominent, we can use multiple square brackets to further decrease their effect:

(Cat) in (forest) at (night) [[blurry lights]]

With these simple adjustments, all Rochatters can control the relative importance and prominence of each element in the generated image.

Give it a try~💙🧡🩵🩷💛💜