InvokeAI: BEST WebUI for Stable Diffusion? - I`m in LOVE!!!

Olivio Sarikas
1 Dec 202211:21

TLDRInvoke AI is presented as one of the best web interfaces for stable diffusion, offering an intuitive setup and user experience. The platform supports Windows, Mac, and Linux, requiring as little as 4GB of RAM. It provides a text-to-image mode, image-to-image mode, and a powerful unified canvas for out-painting, which allows users to achieve high-quality results. The interface is equipped with a variety of settings and tools, such as phase restoration, upscaling, and a brush for color painting. Users can also save their work to the gallery, copy it to the clipboard, or download it. The platform is consistently maintained across restarts and offers hotkeys for efficiency. The community is active and supportive, with an official Discord channel for assistance. Overall, Invoke AI is an impressive tool for image generation and editing, with a user-friendly interface and extensive functionality.


  • 🌐 Invoke AI is a highly intuitive and user-friendly web UI for Stable Diffusion that works across Windows, Mac, and Linux platforms.
  • 💻 It is designed to be compatible with GPUs with as little as 4 gigabytes of RAM, making it accessible to a wide range of users.
  • 📥 The setup process is straightforward, involving downloading an install script and following the on-screen instructions.
  • 🔄 After installation, the system downloads necessary models and organizes them within the Invoke AI folder.
  • 🖼️ The interface includes a text-to-image mode, an image-to-image mode, and a unified canvas for advanced image manipulation.
  • 🎨 The out-painting feature allows users to extend images with high-quality results, even on the first attempt.
  • 📝 A note mode is planned for future updates, which will enable the connection of different modes to generate images.
  • 🛠️ Post-processing capabilities and training options for textual inversion and DreamBooth are also in the pipeline.
  • ⚙️ Users can easily adjust settings such as CFG scale, width, height, and samplers with explanations provided for each.
  • 🖱️ The interface offers tools like phase restoration, upscaling, and a viewer for a clearer image inspection.
  • 🔄 The gallery feature maintains consistency across sessions, allowing users to pick up from where they left off.
  • 📘 Comprehensive settings and options are available for fine-tuning the image generation process to suit individual needs.

Q & A

  • What is InvokeAI and why is it considered one of the best web UIs for Stable Diffusion?

    -InvokeAI is a web interface designed for Stable Diffusion, which is highly intuitive and easy to set up. It is considered one of the best due to its user-friendly design, extensive functionality, and compatibility with multiple operating systems including Windows, Mac, and Linux, even on GPUs with as little as 4 gigabytes of RAM.

  • How can one download and install InvokeAI?

    -To download and install InvokeAI, you need to go to the page where the install script is available. There are install scripts for Linux, Mac, and Windows. After downloading the script, you unpack the zip file and copy the 'invoke AI' folder to your desired location on your hard drive. Then, run the file 'wind lock paths enable' to enable longer paths on Windows and add a new entry to the registry. After that, click on 'install bet' to open the command line and let the installation process begin.

  • What are the different modes available in InvokeAI for image generation?

    -InvokeAI offers several modes for image generation including a text-to-image mode, an image-to-image mode, and a unified canvas for out-painting. It also plans to add a note mode for connecting different modes, a post-processing mode, and training capabilities for textual inversion and DreamBooth.

  • How does the text-to-image mode work in InvokeAI?

    -In the text-to-image mode, users can input a prompt in the prompt area. If a negative prompt is needed, it can be enclosed in square brackets. The 'Invoke' button will then render the images based on the prompt. Users can also set the number of images, steps, CFG scale, width, height, and choose from different samplers.

  • What is the purpose of the unified canvas in InvokeAI?

    -The unified canvas in InvokeAI is used for out-painting, allowing users to extend and create variations of an image. It provides a large, scalable workspace that can be zoomed in and out, and it maintains consistency of the original image while allowing for extensive creative control.

  • How can users interact with the generated images in InvokeAI?

    -Users can interact with the generated images by sending them to the unified canvas, copying the local link, or downloading the image to save it on their drive. They can also use the viewer to zoom in and out for a better view, and make edits using tools like face restore, upscaling, and phase restoration.

  • What kind of support and community does InvokeAI offer?

    -InvokeAI offers a supportive community through its official Discord channel where users can get help from the team and other community members. There is also a GitHub link for reporting bugs and suggesting improvements.

  • How does InvokeAI handle the consistency of the user interface after a restart?

    -InvokeAI maintains consistency of the user interface even after a restart, unless the cache is deleted. This means that the interface will remember the settings and images from the last session, providing a seamless user experience.

  • What are the hotkeys in InvokeAI for faster operation?

    -InvokeAI includes hotkeys to speed up the interface operation. While the specific hotkeys are not detailed in the transcript, they are designed to enhance the workflow and efficiency of users.

  • What are the additional settings and tools available in InvokeAI for image editing?

    -InvokeAI provides a variety of tools for image editing, including a brush for painting and masking, an eraser tool, an infill tool, and options to merge visible parts, save to the gallery, copy to clipboard, and download the image. It also offers unlimited undo steps, allowing users to revert changes as needed.

  • How does InvokeAI assist in maintaining high-quality results during out-painting?

    -InvokeAI assists in maintaining high-quality results during out-painting by allowing users to set a higher scale for the rendering box in the 'infill and scaling' section. This ensures that the box is rendered at a higher resolution and then scaled down, preserving the quality of the extended image.

  • What are the steps to perform out-painting using InvokeAI?

    -To perform out-painting, users should first select the move tool mode (V on the keyboard), then set the rendering area by dragging the square to the desired location and resizing it using the handles. It's recommended to have a lot of overlap for the AI to understand the context. After setting the area, users can click 'invo' to start the rendering process, and can either accept or discard the result. Multiple images can be rendered simultaneously, and users can navigate between different versions.



🚀 Introduction to Invoke AI: A User-Friendly Web UI for Stable Diffusion

The video introduces Invoke AI, a web user interface designed for stable diffusion that is easy to set up and intuitive to use. It can run on Windows, Mac, and Linux, and is suitable for GPUs with as little as 4 gigabytes of RAM. The presenter guides viewers through the installation process, which involves downloading an install script, unpacking a zip file, and running a file to enable longer paths on Windows. The interface includes a text-to-image mode, an image-to-image mode, and a unified canvas for out painting, which the presenter praises for its effectiveness. The video also mentions upcoming features like a note mode, post-processing mode, and training options for textual inversion and dream Booth. The presenter encourages viewers to try out the interface and join the official Discord for support.


🖼️ Exploring Invoke AI's Features: Image Editing and Canvas Manipulation

The presenter delves into the features of Invoke AI, showcasing its image editing capabilities. Viewers are shown how to use the viewer to zoom in and out for a better look at the image, and how to switch between editing modes. The video explains the use of prompts, negative prompts, and the various settings available for image generation, such as the number of images, steps, CFG scale, width, height, and samplers. The presenter also discusses post-creation adjustments like phase restoration and upscaling, and the ability to send images to the unified canvas or download them. The gallery feature is highlighted for its consistency across sessions, and the video demonstrates how to switch between different models and use hotkeys for efficiency. The unified canvas is explored in depth, showing how users can create large renders and use tools like a brush, eraser, and infill. The presenter emphasizes the importance of setting a higher scale for smaller boxes to maintain image quality and suggests using the original colors from the image for a more cohesive composition.


📘 Final Thoughts and Additional Tools in Invoke AI's Web UI

The video concludes with the presenter's admiration for Invoke AI's web UI and an invitation for viewers to try it out. The presenter outlines additional tools and settings available in the UI, such as merging visible parts, saving to the gallery, copying to the clipboard, and downloading images. The presenter also discusses the unlimited undo feature, which allows users to revert changes as far back as their computer's capacity allows. The video provides guidance on uploading images for in-painting and accessing specific settings for the unified canvas. The presenter encourages viewers to clear the canvas history if their computer becomes slow after long work sessions. The video ends with a call to action for viewers to like the video if they enjoyed it and hints at other content they might find interesting, reiterating the invitation to leave a like and the hope to see viewers again soon.



💡Stable Diffusion

Stable Diffusion refers to a type of machine learning model used for generating images from textual descriptions. It's a significant theme in the video as the InvokeAI web interface is designed to work with this technology, allowing users to create images based on prompts.

💡Invoke AI

Invoke AI is a new web interface for interacting with Stable Diffusion models. It is central to the video as the presenter discusses its ease of setup, intuitive use, and various features that facilitate image generation.

💡Web UI

Web UI stands for Web User Interface, which is the platform through which users interact with the software. In the context of the video, Invoke AI's Web UI is highlighted for its user-friendly design and functionality in creating and editing images.

💡Text to Image Mode

This mode allows users to generate images based on textual prompts. It is a key feature of the Invoke AI interface, showcased in the video as a straightforward way to create images from descriptions.

💡Image to Image Mode

This feature enables the editing or transformation of existing images into new ones. It is depicted in the video as a practical tool within the Invoke AI interface for users who want to modify or enhance their images.

💡Unified Canvas

The Unified Canvas is a workspace within the Invoke AI interface that allows for extensive image editing, including outpainting to extend the boundaries of an image. It is presented as a powerful and unique feature of the platform.


Outpainting is the process of generating additional image content beyond the edges of the original image. It is an important concept in the video, demonstrating the capability of the AI to expand images in a coherent manner.

💡CFG Scale

CFG stands for 'Control Flow Graph', and in the context of Stable Diffusion, the CFG scale determines how closely the generated image adheres to the textual prompt. It is a crucial setting within the Invoke AI interface for controlling the output's fidelity to the input prompt.


Samplers are algorithms used in the process of generating images from the Stable Diffusion model. They are referenced in the video as methods that determine the approach taken to create the final image, with different samplers potentially leading to varied results.

💡Phase Restoration

Phase Restoration is a feature that allows for the improvement of image quality, particularly in areas that may have been distorted or lost in the initial rendering process. It is mentioned in the video as an option within the Invoke AI interface for post-image generation refinement.

💡Discord Community

The Discord Community refers to an online forum where users can discuss, get help, and share experiences related to Invoke AI. It is emphasized in the video as a valuable resource for users to connect with the developers and other users for support and collaboration.


Invoke AI is a user-friendly web UI for stable diffusion that is easy to set up and intuitive to use.

The interface is designed to run on Windows, Mac, and Linux, and is compatible with GPUs with as little as 4 gigabytes of RAM.

Invoke AI can be downloaded and installed using provided scripts for different operating systems.

The installation process involves unpacking a zip file and running a setup script that guides users through configuration.

The UI automatically downloads necessary models for use within the interface.

Invoke AI features a text-to-image mode and an image-to-image mode, both of which are highly intuitive.

A unified canvas allows for outstanding out-painting results with minimal effort.

The interface will include a note mode for connecting different modes to generate images.

Post-processing mode and training features for textual inversion and dream Booth are upcoming features.

Users can input prompts and negative prompts to guide the image generation process.

The interface provides controls for the number of images, steps, CFG scale, width, height, and samplers.

Settings are explained through tooltips for user clarity, enhancing the intuitiveness of the interface.

Images can be sent to the unified canvas, copied as a link, or downloaded for personal use.

A viewer mode allows users to inspect images without the distraction of settings.

The interface maintains consistency across restarts, remembering the user's last session.

Hotkeys are available for faster navigation within the interface.

Users can report bugs and access a supportive community through an official Discord link.

The in-paint mode offers a variety of tools for creating variations or rendering from an image.

The unified canvas is highly customizable, allowing for extensive out-painting and image manipulation.

Users can adjust the resolution and quality of the rendered areas for better image results.

The interface includes a comprehensive set of tools for painting, masking, and infill, as well as options for merging, saving, and undoing actions.