Krita AI Diffusion is a free, open source plugin that brings Stable Diffusion, SDXL and Flux into Krita, the open source painting application. It adds a docker where you type a prompt and generate into a layer, inpaint a selection, extend the canvas, upscale, or use Live mode, which repaints the picture as you draw. Under the hood it drives a ComfyUI server, which the plugin can install for you or connect to remotely.

Perfect for artists who already paint in Krita or Photoshop-like tools and want AI as a brush, not a separate website. It needs a decent GPU locally or a ComfyUI server elsewhere.

No install required - run AI models on the cloud

BitVector Prism is the easiest way to start: pick a model, type a prompt and generate in a clean web app, with nothing to configure. BitVector is also available on Discord and on the web (SpyGlass).

Try BitVector

What you need

Krita 5.2 or later on Windows, Linux or macOS. For local generation: an NVIDIA GPU with 6 GB VRAM or more (12 GB for Flux); AMD on Linux with ROCm; Apple Silicon Macs work with MPS. Around 10 GB for the managed ComfyUI server plus models.

How to get and install Krita AI Diffusion

Install Krita from krita.org if you do not have it.
Download the plugin ZIP from the GitHub releases page. In Krita, choose Tools > Scripts > Import Python Plugin from File and pick the ZIP, then restart Krita.
Enable the plugin under Settings > Configure Krita > Python Plugin Manager if it is not already, and open the AI Image Generation docker (Settings > Dockers).
In the docker, choose Local Managed Server and press Install: the plugin downloads ComfyUI and the required models (several gigabytes). Alternatively, connect to an existing ComfyUI by address.
Pick the model packages you want (SD 1.5, SDXL, Flux) in the server settings; you can also point the plugin at an existing models folder.

Download only from the official links on this page. Installers found elsewhere may be out of date or tampered with. ModelTrends.ai hosts no files.

No install required - run AI models on the cloud

PirateDiffusion is Telegram only and built for pros: thousands of models, LoRAs and workflows driven by chat commands, with unlimited generation on a fixed price plan.

Try PirateDiffusion

How to use Krita AI Diffusion

Create a new document, write a prompt in the docker and press Generate. The result appears as a new layer; keep the version you like.
Make a selection and press Generate to fill or change only that region (inpainting), using your prompt and the surrounding image.
Enlarge the canvas (Image > Resize Canvas), select the empty area and generate to outpaint.
Switch to Live mode, pick a strength and start painting rough shapes: the AI repaints them in real time as you draw.
Open the style settings to choose the checkpoint, add LoRAs with weights (and their trigger words from the model page) and set the sampler.
Use Control layers (pose, depth, canny, scribble, reference) to guide the result from another layer of your document.

Tips and common problems

Styles let you save a checkpoint, LoRAs and prompt prefixes under a name; create one per look you use often.
If Krita cannot connect to the server, check the server tab in the docker for the log; most problems are a model download that did not finish.
Live mode is fastest with SD 1.5 or a lightning SDXL model; set a lower resolution while sketching and upscale at the end.
You can use a ComfyUI running on another PC or a rented GPU: set its address in the connection settings.

Models you can run in Krita AI Diffusion

Krita AI Diffusion runs Stable Diffusion 1.5, SDXL and Flux checkpoints and LoRAs from this catalog, plus the ControlNet and IP-Adapter models it installs. Video models are not supported.

Flux Models

Stable Diffusion XL 1.0 Models

Stable Diffusion 1.5 Models

No install required - run AI models on the cloud

No GPU, no downloads, no Python. These services run the same open source models on their own servers, so you can start generating from a browser, Discord or Telegram in a minute.

Try BitVector

Try PirateDiffusion

Links

These links go to the makers' own sites. ModelTrends.ai is not affiliated with them and does not host any software.

Other local apps

How to install ComfyUI

The node-based powerhouse for Stable Diffusion, Flux, Wan video and almost every new open source model. Free, open source, for Windows, Mac and Linux.