Repo Image & Video Limited info

Acly/krita-ai-diffusion

Krita plugin for generative AI: inpainting, live painting, upscaling, ControlNet and regions, running locally via ComfyUI or in the cloud.

  • 10.7k GitHub stars
  • Python
  • ⚖️ GPL-3.0
  • 🎯 Intermediate
View on GitHub (opens in a new tab) Setup guide
Acly/krita-ai-diffusion preview image

What it is

Krita AI Diffusion is a plugin that brings generative AI into image painting and editing workflows inside Krita. It supports inpainting with selections, live painting, upscaling, edit models, ControlNet, IP-Adapter and regions. It uses ComfyUI as its backend and can run locally on your hardware or via cloud generation.

Who it's for

  • Krita users who want AI generation integrated into their painting and editing workflow
  • Artists who want control over results through selections, regions, reference images, sketches, line art and depth maps
  • Users who prefer local, open-source models and customizable presets

Requirements

Requirements

  • Krita 5.2.0 or newer
  • Windows, Linux or macOS
  • For local use, a powerful graphics card with at least 6 GB VRAM (NVIDIA) is recommended
  • GPU support: NVIDIA via CUDA, AMD via ROCm, Intel via XPU (Windows/Linux), Apple Silicon via MPS on macOS 14+
  • CPU is supported but very slow

Setup

  1. Install Krita

    Install Krita if you haven't yet. Required version: 5.2.0 or newer.

  2. Download the plugin

    Download the plugin ZIP archive from the latest release page: https://github.com/Acly/krita-ai-diffusion/releases/latest

  3. Import the plugin into Krita

    Start Krita and choose Tools ▸ Scripts ▸ Import Python Plugin from File..., then point it to the downloaded ZIP archive.

  4. Restart and show the docker

    Restart Krita and create a new document or open an existing image. To show the plugin docker, go to Settings ‣ Dockers ‣ AI Image Generation.

  5. Configure the server

    In the plugin docker, click "Configure" to start local server installation or connect.

Pros & cons

Pros

  • Pro:Selection-based inpainting and expand give precise control over where generation happens
  • Pro:Wide feature set: live painting, upscaling, ControlNet, IP-Adapter, regions, edit models, job queue and history
  • Pro:Runs locally with open models, with cloud generation available as an alternative
  • Pro:Customizable presets (checkpoints, LoRA, samplers) alongside strong defaults for a streamlined UI

Cons

  • Con:Local generation needs a powerful GPU (at least 6 GB VRAM NVIDIA recommended); CPU is very slow and low memory may cause failures
  • Con:Requires Krita 5.2.0 or newer
  • Con:Segmentation/object selection tools are not included and require a separate plugin

Images

Acly/krita-ai-diffusion preview image
Acly/krita-ai-diffusion preview image · Acly/krita-ai-diffusion on GitHub · GPL-3.0
Acly/krita-ai-diffusion preview image