Midjourney On The Web - All You Need To Know | Midjourney Alpha

AI Tooltip
9 May 202423:00

TLDRThe video script introduces the new web interface for Midjourney, which simplifies the process of generating images without memorizing codes. It now offers a user-friendly UI accessible to all, featuring tools from the Discord version. The host demonstrates how to use various parameters for image generation, including model selection, aspect ratio, and stylization, and provides tips for accurate text rendering. The video also covers post-generation image manipulation tools and concludes with a guide on generating text within images using specific prompts.


  • ๐Ÿ˜€ Midjourney has transitioned from memorizing codes to a user-friendly web interface, making it accessible to anyone with over 100 generations in their account.
  • ๐ŸŒ The new UI for Midjourney is browser-based, allowing users to easily access and use the tool without the hassle of memorizing commands.
  • ๐Ÿ”‘ Users with fewer than 100 generations can generate more to gain access to the web platform, which was initially limited to those with a higher count but has now been made more widely available.
  • ๐Ÿ› ๏ธ The UI has been significantly improved over the past year, integrating most of the tools previously available only in the Discord version of Midjourney.
  • ๐Ÿ“š The video aims to educate viewers on how to use each tool within the UI, as the platform itself does not provide detailed explanations.
  • ๐ŸŽจ Midjourney offers two models: the standard model for aesthetically pleasing images and the raw model which relies entirely on the user's prompt for styling.
  • ๐Ÿ“ Users can adjust the aspect ratio of their images using a slider in the UI, with options for square, landscape, and portrait orientations.
  • ๐Ÿ”„ After generating an image, Midjourney provides options to modify the image, such as subtle and strong variations, as well as the ability to upscale the image with subtle or creative changes.
  • ๐Ÿ” The 'remix' feature allows users to modify the original prompt while keeping the changes subtle or making more significant alterations.
  • 't๐Ÿ–ผ๏ธ With the 'Very Regent' feature, users can select an element in the image and request Midjourney to modify it, such as changing a squirrel to a cat or adding an explosion.
  • ๐Ÿ“ To generate accurate text within images, users should include the desired text within double quotations in their prompt, and using the raw model or reducing stylization can improve text accuracy.

Q & A

  • What is the main change in the new UI for using Midjourney?

    -The main change is that Midjourney can now be used through a browser interface, making it more accessible and user-friendly, similar to other tools. It's no longer necessary to memorize codes, and users can set up parameters with just a few clicks.

  • Who has access to the new Midjourney web platform?

    -Initially, it was released to users with over 100 generations in their account, but the threshold has been lowered, making it publicly available for everyone.

  • What improvements have been made to the Midjourney UI over the past year?

    -The UI has been modified and improved to be more usable and visually appealing. It now includes most of the tools previously available only in Midjourney on Discord.

  • What are the two different models available in Midjourney for image generation?

    -There are two models: the standard Midjourney model that creates aesthetically pleasing images, and the raw model that relies solely on the prompt input and does not apply any styling on its own.

  • How can users ensure they get more accurate and realistic images in Midjourney?

    -Users can describe the scene, shot, lighting, effects, and mood in detail in their prompt and choose the raw model with zero stylization for more realistic and accurate images.

  • What does the 'Weirdness' scale in Midjourney do?

    -The 'Weirdness' scale controls the unusualness of the generated images compared to the average output of Midjourney, resulting in more quirky and unexpected results as the scale is increased.

  • What is the purpose of the 'Variety' option in Midjourney?

    -The 'Variety' option increases the difference between the four generated images, providing a range of styles and compositions for users to choose from, especially when they are unsure about the style they want.

  • How can users manipulate images after they have been generated in Midjourney?

    -Users can use tools like 'Remix' to modify the prompt, 'Upscale' to increase resolution, 'Pan' to add additional areas to the image, 'Zoom' to provide surroundings, and 'Very Regent' to select and modify specific elements in the image.

  • What is the trick to include accurate text in the generated images by Midjourney?

    -To include text, it must be placed within double quotations in the prompt. Midjourney recognizes double quotations as text to be included in the image.

  • How can users get more accurate text rendering in Midjourney?

    -To get more accurate text rendering, users can reduce the stylization or use the raw model. This will help Midjourney to better understand and render the text as intended.

  • What additional resources are provided by the video creator to help users understand and use Midjourney?

    -The video creator provides a detailed cheat sheet with all the information discussed in the video, as well as a newsletter subscription for weekly content on AI tools, tips, and tricks.



๐ŸŒ New UI for Mid Journey: Browser-Based Access

The video script introduces a new user interface for Mid Journey, a tool previously requiring memorization of complex codes. The updated UI is now accessible through a browser, making it more user-friendly and widely available to users with over 100 generations in their account. The script highlights the lowered threshold for access and the inclusion of most tools previously only available in Discord. The narrator promises to demonstrate the functionality of each tool in the UI, focusing on its usability and the absence of user-friendly documentation.


๐ŸŽจ Exploring Image Generation with Mid Journey's Standard and Raw Models

This paragraph delves into the comparison between Mid Journey's standard and raw models for image generation. The standard model automatically applies styling to create aesthetically pleasing images, whereas the raw model relies solely on the user's prompt for style, offering greater accuracy for those who specify their desired look. The script discusses the importance of accurate description for the raw model and demonstrates the impact of stylization levels on image creativity and realism.


๐Ÿ”„ Post-Generation Image Manipulation Features in Mid Journey

The script explains the various post-generation tools available in Mid Journey for manipulating images. These include the ability to upscale images while maintaining or enhancing details, remix images with subtle or strong deviations from the original, and the pan feature that expands the image in a specified direction. Additionally, the zoom option is introduced, which provides a broader view of the image surroundings, and the rerun feature, whichๅ†็”Ÿs the same prompt for varied outcomes.


๐Ÿ›  Advanced Image Editing with Mid Journey's 'Very' Region Tool

This section introduces the 'Very' region tool in Mid Journey, which allows users to select and modify specific elements within an image. The tool enables the replacement of elements with others, such as changing a squirrel to a cat or adding an explosion within the scene. The script emphasizes the creative potential of this feature for generating unique and customized images.


๐Ÿ“œ Accurate Text Rendering in Mid Journey Images

The final paragraph focuses on the capability of Mid Journey to render text accurately within generated images. It outlines a specific method for including text within prompts, using double quotations, and suggests reducing stylization or using the raw model for better text accuracy. The script demonstrates the effectiveness of this technique with various examples and discusses the potential for text-based image generation.




Midjourney refers to a specific tool or platform that allows users to generate images based on textual prompts. In the video, it is mentioned as having transitioned from a system requiring memorization of codes to a more user-friendly web-based interface. The script discusses various features and parameters that can be adjusted within Midjourney to customize the image generation process.

๐Ÿ’กUI (User Interface)

The user interface, or UI, is the space where interactions between the user and the system occur. In the context of the video, the UI of Midjourney has been improved and is now accessible through a browser, allowing users to generate images with a more intuitive and streamlined process.


Parameters are the specific settings or options that users can adjust to influence the outcome of the image generation in Midjourney. The script describes how users can set parameters such as model type, aspect ratio, and stylization to customize their images according to their preferences.


In the context of Midjourney, a model refers to the underlying algorithm or system that interprets the user's prompt and generates an image. The video script mentions 'standard' and 'raw' models, with the standard model applying its own styling and the raw model relying solely on the user's prompt for styling direction.


Stylization in Midjourney determines the degree to which the generated image will exhibit a particular style or aesthetic. The script explains that higher stylization values lead to more creative and dramatic images, while lower values result in more accurate and less embellished outcomes.

๐Ÿ’กWeirdness Scale

The weirdness scale is a parameter in Midjourney that controls the unusualness or uniqueness of the generated images. The video describes how adjusting this scale can lead to more quirky and unexpected results, deviating from the average output of the platform.


Variety, as discussed in the script, is a parameter that influences the diversity of the four images generated by Midjourney. A higher variety setting results in more distinct and different images, offering users a range of options to choose from.


Upscale is a feature within Midjourney that increases the resolution of an image. The script distinguishes between 'upscale subtle', which maintains the original structure, and 'upscale creative', which adds more elements and changes to the image while enhancing the resolution.


The remix feature in Midjourney allows users to modify the original prompt of a generated image while keeping the overall theme or environment consistent. The script explains that there are subtle and strong remix options, with the strong option resulting in more noticeable changes to the image.

๐Ÿ’กVery Region

Very Region is a tool within Midjourney that enables users to select a specific element in an image and request modifications to it. The video script illustrates how this feature can be used to change elements like turning a squirrel into a cat or adding an explosion to a scene.

๐Ÿ’กText Rendering

Text rendering in Midjourney refers to the platform's ability to interpret and visually represent text described in the user's prompt. The script provides a trick for including accurate text by using double quotations and suggests reducing stylization or using the raw model for better text accuracy.


Midjourney now offers a web-based UI, eliminating the need to memorize complex codes.

The new UI is accessible to users with over 100 Generations in their account, making it more inclusive.

Over the past year, the UI has been significantly improved for usability.

Most of the tools previously available on Discord have been integrated into the new web UI.

The video will demonstrate each tool's functionality and optimal use cases.

A clean and precise guide to using Midjourney tools is provided, simplifying the process.

The video is divided into three parts: pre-generation settings, post-generation image manipulation, and tips for maximizing Midjourney.

The new interface includes an 'Imagine' box for browsing creations by others for inspiration.

Users can organize their images in various ways, including archiving and creating folders.

Parameters can be set before generating an image, including model selection and aspect ratio.

The 'Raw' model relies on user-provided style descriptions for image generation.

Stylization control allows for more creative or realistic image outcomes.

The 'Weirdness' scale introduces more unusual and unique elements into the images.

The 'Variety' option ensures diverse outcomes among the four generated images.

The 'Speed' panel offers options for faster generation at the cost of GPU time.

The 'Plus' feature enables users to upload images for style or character reference.

Post-generation tools allow for modifications such as variations, upscaling, remixing, panning, and zooming.

The 'Very' feature permits element-specific modifications within an image.

Accurate text rendering in images can be achieved using a specific prompt format.

A comprehensive guide and tips for crafting effective prompts are provided for reference.