Plugin Installation

This is a step-by-step guide for downloading and installing the Generative AI plugin for Krita.

Installing the Plugin

1. Make sure you have Krita version 2.2 (or later) installed.

   You can download and install it from krita.org .

2. Download the Plugin

   Download the latest release of the Plugin. Remember where you save the ZIP file.

3. Install the Plugin

   Run Krita. Go to Tools ▸ Scripts ▸ Import Python Plugin from File…

   

   Select the ZIP file you just downloaded. When prompted to enable the Plugin, press Yes.

   

   Now Krita needs to be restarted!

4. Enable the Docker

   In Krita, create a new Document. For starters, choose a low resolution.

   

   Enable the Plugin Docker from the Menu bar: Settings ▸ Dockers ▸ 🗹 AI Image Generation

   

5. Configure Image Generation

   The Plugin is installed now and you should see the Docker appear.

   

   To generate images you also need a backend server. Click Configure to choose between multiple options:

   

   • Local Managed Server

     The Plugin will download and install the server on your system.
     Requires a powerful GPU! Continue reading below for details.

   • Online Service

     Create an account and generate images on cloud GPU workers.
     This requires no installation or hardware on your end. Pricing and Terms

   • Custom Server

     The Plugin can connect to an existing ComfyUI server, either local or remote.

     Read about required custom nodes and models here. This is the most flexible option, but some technical knowledge is required.

     Caution

     If you choose this option, it is your responsibility to install and run the server. Instructions and tutorials for ComfyUI can be found online.

6. Connect to Online Service (Optional)

   Press the Sign in button. This should open your browser, where you need to log in, or sign up for an account.

   

   Once done, you should see the confirmation screen and can start generating images!

   

Troubleshooting

Please check Common Issues and search for existing issues .

If you cannot find a solution to your problem, open a new Issue .

For questions, reach out via Discussions or join us on Discord .

Caution

Official Krita channels cannot offer support for issues related to the plugin.

Local Server Installation

These steps are only required if you selected Local Managed Server as the backend.

1. Hardware requirements

   NVIDIA	GTX 1060 6GB or better. You may encounter issues and limitations on older cards.
   AMD	12GB VRAM or more. Slow with limited feature set. Custom install required for Linux.
   Apple	Apple Silicon via MPS (community support).
   CPU	Works, but is very slow.

   Please make sure you are using an up-to-date GPU driver!

2. Operating System

   Windows, Linux or macOS

   Linux and macOS users

   Python with venv must be installed. Recommended version: 3.11 or 3.12.
   This is usually available via package manager, eg. apt install python3-venv

   Linux users

   Some packages cannot be installed while running the Flatpak version of Krita. It is recommended to use the AppImage instead.

3. Installation

   1. Choose a path where to install.
      Keep in mind AI models can be very large, 10GB is a minimum, 50GB+ not unusual depending on optional packages.
   2. Choose a backend. CUDA only works for NVIDIA GPUs. Use DirectML for AMD. Make sure to do this before you install!
   3. Select packages. You need at least one workload and one checkpoint.
      Everything else is optional and will considerably increase download size. You can come back later and add packages at any time.
   4. Click install!