Tokens are the individual components that make up your input prompt. They can be as simple as common words, which each get their own token, or as complex as less common words and special characters, which might get split into multiple tokens.

• Common words: Words frequently used in language (like 'the', 'and', 'a', etc.) typically have their own unique token.

• Complex/less common words: Words that are less frequently used or more complex might be broken down into multiple tokens.

• Special characters and emojis: These often get split into multiple tokens as well.

Any tokens beyond the 74 token limit are ignored by midjourney. This is important to keep in mind to ensure your entire prompt is considered during image generation.

However, with multi-prompts (prompts separated by '::'), each prompt has its own 74 token limit, giving you greater influence over the resulting image.

This token limit isn't often discussed due to the complexity of tokens and the lack of a straightforward conversion rate from words to tokens. This tool helps to visualize and understand how your prompts are broken down into tokens, and what is being ignored without your knowledge if you exceed the invisible token limit.

The Tokens List displays all the tokens that make up your prompt along with their corresponding IDs. These IDs are the internal values that are actually passed to the network.

This tool helps visualize the tokenization process, providing you with greater insight into how the model interprets your prompts.

Prompts are the text inputs you provide to guide the image generation. A multi-prompt is a technique where multiple prompts are used, separated by ' :: '. The intention is to provide more granular control over different aspects of the image.

Each prompt within a multi-prompt has its own limit of 74 tokens. This means that each separated prompt can influence the image independently without being constrained by the token count of other prompts.

This tool fully supports multi-prompts. When calculating token and word counts, the tool intelligently identifies and does not include the multi-prompt separators (' :: ') in the count. This ensures an accurate representation of your input for the image generation process.

• Single Prompt: "A sunny beach with palm trees"

• Multi-Prompt: "A sunny beach :: Palm trees on the right :: Clear blue water"