DEV.co
AI Frameworks · Acly

krita-ai-diffusion

Krita AI Diffusion is a Krita plugin enabling AI-powered inpainting, outpainting, and image generation directly within the editor. Users can refine existing artwork with text prompts and ControlNet guidance while maintaining precise control over generation regions and strength.

Source: GitHub — github.com/Acly/krita-ai-diffusion
10.3k
GitHub stars
600
Forks
Python
Primary language
GPL-3.0
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
RepositoryAcly/krita-ai-diffusion
OwnerAcly
Primary languagePython
LicenseGPL-3.0 — OSI-approved
Stars10.3k
Forks600
Open issues110
Latest releasev1.52.1 (2026-06-30)
Last updated2026-06-30
Sourcehttps://github.com/Acly/krita-ai-diffusion

What krita-ai-diffusion is

Python-based Krita plugin leveraging ComfyUI as a diffusion backend, supporting Stable Diffusion, Flux, and ControlNet models with local GPU acceleration (NVIDIA/CUDA, AMD/ROCm, Intel/XPU, Apple/MPS) or cloud fallback. Integrates IP-Adapter for style transfer and region-based multi-prompt generation.

Quickstart

Get the krita-ai-diffusion source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/Acly/krita-ai-diffusion.gitcd krita-ai-diffusion# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Iterative Digital Painting & Asset Refinement

Artists refining hand-painted artwork, adding/removing objects, and upscaling compositions to 4K/8K within a familiar Krita workflow without switching tools.

Concept Art & Design Iteration

Design teams using ControlNet (scribble, pose, depth, line art) to explore variations on sketches and concepts while maintaining compositional control via selections and reference images.

Content Production for Games & Animation

Studios leveraging inpainting, outpainting, and upscaling to accelerate texture/background generation and refinement at scale, with optional cloud generation for resource-constrained environments.

Implementation considerations

  • ComfyUI backend must be installed locally or a remote server configured; automatic setup available but may fail on atypical system configurations—FAQ and Discord support required for troubleshooting.
  • GPU drivers and libraries (CUDA/ROCm/XPU/MPS) must be pre-installed and compatible; plugin does not bundle or validate driver state, only detects supported hardware at runtime.
  • Model downloading is on-demand (Flux 2, Illustrious, Stable Diffusion variants); first-run can be slow and requires significant disk/network bandwidth; no built-in quota or caching strategy documented.
  • Krita version pinned to 5.2.0+; verify compatibility matrix as plugin evolves; official Krita support channels explicitly exclude this plugin (direct community support via Discord/GitHub).
  • Optional segmentation features require a separate companion plugin (krita-ai-tools) for object selection; LoRA and custom checkpoint support documented but requires manual model management.

When to avoid it — and what to weigh

  • No GPU / CPU-only deployment — Local CPU-based generation is explicitly noted as 'very slow.' Requires at least 6GB VRAM (GPU) for practical performance; cloud fallback available but adds latency and potential cost.
  • Closed-source / Proprietary AI pipelines required — Plugin is built on open-source models (Stable Diffusion, Flux, ComfyUI). If vendor-locked or proprietary diffusion models are mandated, this will not satisfy requirements.
  • Standalone headless/API-only use case — Tightly coupled to Krita UI and workflow; not designed as an independent inference service or REST API. Requires Krita 5.2.0+ as a hard dependency.
  • Zero tolerance for GPL licensing — GPL-3.0 requires derivative works and linked codebases to also adopt GPL-3.0. Commercial products embedding this plugin must meet GPL-3.0 obligations; internal/non-redistributed use is permissible.

License & commercial use

GPL-3.0 (GNU General Public License v3.0). This is a copyleft license requiring all derivative works, modifications, and linked code to be distributed under GPL-3.0 and made publicly available with source code. Internal use without redistribution is permitted; commercial use is permissible but only if source and license obligations are met.

Commercial use is legally permissible under GPL-3.0 if (1) any modifications or derivative works are published under GPL-3.0 with source code, and (2) end-user rights to access/modify source are preserved. Embedding this plugin in a closed-source commercial product without GPL-3.0 compliance is non-compliant. Consult legal counsel if commercial distribution is intended. Internal use within a company (non-distributed) is allowed.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationStrong
License clarityClear
Deployment complexityModerate
DEV.co fitGood
Assessment confidenceHigh
Security considerations

Local GPU execution avoids cloud data transmission, reducing exposure of user images if deployed on-premises. ComfyUI backend runs locally by default but network-exposed if configured for remote access (e.g., across a company LAN); no built-in authentication or encryption for remote servers—network security is user's responsibility. Cloud generation option (provider unknown) introduces third-party data handling risk; terms and retention policies are not disclosed in publicly available sources. Models (Stable Diffusion, Flux) are open-source; no attestation of model integrity or provenance beyond upstream repos. No explicit supply-chain security (e.g., signed releases, SBOMs) mentioned.

Alternatives to consider

Automatic1111 WebUI (standalone Stable Diffusion UI)

Web-based, browser-agnostic alternative to Krita plugin; broader model ecosystem; no Krita dependency. Trade-off: requires manual asset pipeline integration and lacks native painting workflow.

Clipdrop / Adobe Firefly (cloud-based inpainting)

Commercial SaaS options with proprietary models and no local setup complexity. Trade-off: vendor lock-in, per-use costs, and data privacy concerns; no open-source models.

ComfyUI standalone with custom frontends (e.g., Invoke, OpenArt)

Direct ComfyUI usage with alternative UIs; maximum flexibility and no Krita coupling. Trade-off: steeper learning curve, no native Krita workflow integration, separate tool ecosystem management.

Software development agency

Build on krita-ai-diffusion with DEV.co software developers

Evaluate krita-ai-diffusion for concept art, asset refinement, and design iteration. Assess GPU requirements, ComfyUI deployment complexity, and GPL-3.0 licensing implications. Contact our team to review integration, security, and commercial use considerations.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

krita-ai-diffusion FAQ

Can I run this entirely offline without cloud services?
Yes. Deploy ComfyUI locally with downloaded models and connect the plugin to localhost (auto-detected). Cloud generation is optional but provided for users without local GPU resources.
What are the minimum GPU requirements?
At least 6 GB VRAM (NVIDIA recommended). AMD (ROCm), Intel (XPU), and Apple Silicon (MPS) are supported. CPU-only mode is supported but 'very slow' per documentation.
Can I use this plugin in a commercial product or game?
Yes, if you comply with GPL-3.0: publish modifications under GPL-3.0 with source code, and preserve end-user rights to access/modify source. Internal non-redistributed use is simpler. Consult legal counsel for commercial distribution.
What models does the plugin support?
Officially supports Flux 2, Z-Image, Stable Diffusion 1.5/XL, Illustrious, and ComfyUI-compatible custom checkpoints/LoRA. Model selection and download are managed via the plugin's configuration UI.

Work with a software development agency

DEV.co helps companies turn open-source tools like krita-ai-diffusion into production software. Our software development services cover the full lifecycle — architecture, web development, integration, and maintenance — delivered by software developers and web developers who ship. Engage our software development agency to implement or customize it for your ai frameworks stack.

Integrate AI-Powered Image Generation into Your Krita Workflow

Evaluate krita-ai-diffusion for concept art, asset refinement, and design iteration. Assess GPU requirements, ComfyUI deployment complexity, and GPL-3.0 licensing implications. Contact our team to review integration, security, and commercial use considerations.