Skip to content
RepoQuarry

Krita AI Diffusion Acly/krita-ai-diffusion

Krita plugin for generative fill, live painting, upscaling and reference-guided generation, using a local or remote ComfyUI backend.

Krita AI Diffusion at a glance

Checked 15 Oct 2026

A plugin that brings image generation into the Krita painting app: inpainting/generative fill and expand, live painting, upscaling to 4k-8k, ControlNet, IP-Adapter style and composition transfer, and regions. It uses ComfyUI as its backend, installed automatically or your own. Models include Flux 2, Z-Image, SD 1.5, SDXL and Illustrious; licence depends on the model. Code is GPL-3.0. 6GB NVIDIA VRAM is recommended locally; an optional paid cloud uses tokens.

Cite this pageRepoQuarry. “Krita AI Diffusion.” Checked 15 Oct 2026. https://repoquarry.dev/projects/krita-ai-diffusion

How much of Adobe Firefly does it replace?

3 of 3 jobs. Closest to Firefly-in-Photoshop generative fill, but inside Krita with open models.

Partial: Text-to-image Partial replacement
Partial: Image editing (generative fill and expand) Partial replacement
Partial: Style and image reference Partial replacement

How much of Midjourney does it replace?

4 of 4 jobs. Editing-first workflow inside Krita rather than prompt-first generation.

Partial: Text-to-image Partial replacement
Partial: Image editing (inpainting and outpainting) Partial replacement
Partial: Upscaling Partial replacement
Partial: Style and image reference Partial replacement

Facts, with sources

Every answer links to where we found it. We read the documentation; we did not install it.

Report a correction
QuestionAnswerSourceChecked
Runs on my machine? Confirmed: Yes, local ComfyUI server; optional paid cloud Customize presets, bring your own models, and run everything local on your hardware. Cloud generation is also available to get started quickly without heavy investment. github.com · krita-ai-diffusion#readme ↗ 1d ago
Code licence Partial: GPL-3.0 (copyleft) github.com · LICENSE ↗ 1d ago
Model weights licence Partial: No single default; recommended checkpoints per style; licence varies by model Checkpoint lists the recommended diffusion models used by the plugin's default styles. Base models: SD1.5, SDXL, Illustrious, Flux, Flux 2 4B, Z-Image. docs.interstice.cloud · models ↗ 1d ago
Commercial use Partial: GPL-3.0 code; weights vary (Flux 2 Klein 4B and Z-Image Turbo are Apache-2.0) FLUX.2-klein-4B and Tongyi-MAI/Z-Image-Turbo are tagged license: apache-2.0 on Hugging Face; FLUX.2-klein-9B is flux-non-commercial-license. huggingface.co · FLUX.2-klein-4B ↗ 1d ago
Paid services needed? Partial: None required; optional cloud tokens (10 EUR for 5,000) Start exploring with a limited number of free tokens. One-time payment 10 EUR for 5,000 tokens. To generate an image at 1024x1024 resolution costs 5 tokens (~0.01 EUR). interstice.cloud · service ↗ 1d ago
Hardware Confirmed: 6GB NVIDIA VRAM recommended; CPU supported but very slow To run locally a powerful graphics card with at least 6 GB VRAM (NVIDIA) is recommended. CPU: supported, but very slow. github.com · krita-ai-diffusion#hardware-support ↗ 1d ago
Operating systems Confirmed: Windows, Linux, macOS 14+ (Apple Silicon MPS) Windows, Linux, MacOS ... Apple Silicon: MPS on macOS 14+ github.com · krita-ai-diffusion#operating-system ↗ 1d ago
Setup effort Confirmed: Easy; import plugin ZIP into Krita 5.2+, server installs from the plugin In the plugin docker, click "Configure" to start local server installation or connect. github.com · krita-ai-diffusion#installation ↗ 1d ago

Get the weekly top 5 and what’s new, every Monday

We’ll email you once a week. Unsubscribe any time. We never share your email.

Cookie preferences

Choose what we may store in your browser. You can change this any time from Cookie settings in the footer.

Essential

Security, and remembering this choice. rq-consent · 12 months

Always on