How to Add More Samplers to the Krita AI Generation Tool

Integrating additional samplers into your Krita AI generation workflow can give you more creative control and generate varied, high-quality images. In this guide, we’ll walk you through everything from understanding what samplers do to step-by-step instructions on how to add custom sampler presets to the Krita AI Diffusion plugin. Whether you’re a digital artist or an AI enthusiast, read on for actionable insights and tips optimized for the latest 2025 trends.

---

Introduction: Krita AI Generation Tool

With the explosion of AI-powered art tools, Krita’s AI Diffusion plugin has become a go-to resource for many artists. By default, the plugin comes with a set of samplers designed to denoise and refine images generated via diffusion models. However, as new techniques emerge, you might want to experiment with additional or custom samplers to improve quality, speed, or creativity. In this blog, we explore:

• What are samplers?
• Why add more samplers?
• Step-by-step guide to adding custom samplers
• Best samplers for AI image generation in 2025
• Troubleshooting tips and FAQs

---

What Are Samplers in AI Image Generation?

Samplers are algorithms that guide the diffusion process by controlling how noise is reduced over successive steps until a final image is revealed. In the context of diffusion models, different samplers such as Euler, DDIM, or DPM++ can drastically influence the visual style and quality of the output. For digital artists, this means that the choice of sampler not only affects the speed of generation but also the artistic characteristics of your final image.

---

Why Add More Samplers?

By expanding your sampler options, you can:

• Enhance Creative Flexibility: Experiment with different image styles and details.
• Improve Generation Quality: Certain samplers converge faster or produce higher-fidelity images.
• Optimize Performance: Balance speed and quality based on your hardware and workflow.
• Customize Workflow: Tailor the AI generation process to suit your artistic vision.

---

Step-by-Step Guide to Adding More Samplers

Below is a detailed walkthrough on how to integrate additional samplers into the Krita AI Diffusion plugin.

Step 1: Update Your Krita AI Diffusion Plugin

1. Download the Latest Release:
   Visit the Acly/krita-ai-diffusion GitHub repository and download the latest ZIP archive.
2. Install/Update the Plugin:
   In Krita, go to Tools > Scripts > Import Python Plugin from File… and select the downloaded ZIP file.
   Tip: Back up your current plugin settings if you plan to experiment.

Step 2: Locate the Sampler Presets File

1. Find the samplers.json File:
   The plugin creates or uses a file (commonly found in <appdata>/krita/ai_diffusion/presets/) that contains existing sampler presets.
2. Open the File:
   Use a text editor (Notepad, VS Code, etc.) to open samplers.json.

Step 3: Create Custom Sampler Presets

1. Review the File Structure:
   The file typically follows a JSON format with each preset defined by its name and parameters (e.g., sampler algorithm, scheduler type, steps, and CFG scale).

2. Add a New Sampler Entry:
   Insert a new JSON object with your desired parameters. For example:{
   “Custom Sampler – DPM++ 3M”: {
   “sampler”: “dpmpp_3m_sde”,
   “scheduler”: “exponential”,
   “steps”: 25,
   “minimum_steps”: 5,
   “cfg”: 7.0
   }
   }

3. Save Your Changes:
   After editing, save the file.

Step 4: Restart Krita to Apply Changes

1. Close and Reopen Krita:
   The plugin only reads the presets at startup, so a restart is required.
2. Verify in the Plugin Docker:
   Go to Settings > Dockers > AI Image Generation and check if your custom sampler appears in the sampler list.

Step 5: Test Your New Sampler

1. Generate an Image:
   Create a new image using a familiar text prompt and select your custom sampler.
2. Evaluate the Output:
   Adjust parameters such as the number of steps or CFG scale if necessary, to fine-tune the performance of the new sampler.

---

Best Samplers for AI Image Generation in 2025

Based on recent trends and user feedback, here are some of the best samplers you might consider:

• Euler / Euler a:
  Ideal for fast convergence and smoother image details.
• DDIM:
  Offers a good balance between quality and generation speed.
• DPM++ 2M / DPM++ 3M:
  Known for high-fidelity image generation with rich details.
• UniPC:
  A new entrant that provides rapid and high-quality sampling in as few as 20–30 steps.
• Heun:
  Provides precise control over noise reduction but might require fewer steps to avoid overprocessing.

Keywords: best samplers 2025, Euler, DDIM, DPM++ 2M, UniPC, Heun, AI image quality

---

Troubleshooting Tips

While adding new samplers is straightforward, you may encounter issues. Here are some common problems and fixes:

• Sampler Not Appearing in the List:
  • Ensure the JSON syntax in samplers.json is correct.
  • Verify that you restarted Krita after saving changes.
• Poor Image Quality or Inconsistent Results:
  • Adjust the “steps” and “minimum_steps” parameters.
  • Try tweaking the CFG (Classifier-Free Guidance) value.
• Plugin Crashes or Freezes:
  • Check for conflicts with other Krita plugins or outdated dependencies.
  • Look into the plugin’s FAQ or GitHub issues page for known problems and fixes.
• Performance Issues:
  • Reduce the number of steps if generation takes too long.
  • Consider hardware limitations (VRAM, GPU model) and use lighter samplers if necessary

FAQ: How to Add More Samplers to the Krita AI Generation Tool

What exactly is a sampler in the Krita AI Diffusion plugin?

Samplers are algorithms that determine how noise is removed during the image generation process. They affect the quality, style, and speed of the output image.

How do I add a custom sampler?

You add a custom sampler by editing the samplers.json file in your plugin’s presets folder, then restarting Krita so the plugin loads your new preset.

Which sampler is best for high-fidelity images?

Many artists favor DPM++ 2M or DPM++ 3M for their balance of detail and consistency, though newer options like UniPC are also highly rated for quality and speed.

What should I do if my custom sampler is not working?

Double-check your JSON syntax, confirm that Krita has been restarted, and consult the plugin’s documentation or GitHub issues page for any compatibility concerns.

Can I switch between samplers during image generation?

Yes, the plugin allows you to select different samplers from the dropdown menu in the AI Image Generation docker, so you can experiment on the fly.

Conclusion

Adding more samplers to the Krita AI generation tool can unlock new creative possibilities and enhance your workflow. By following the simple steps outlined above, you can easily integrate custom sampler presets, experiment with various diffusion algorithms, and achieve higher-quality image outputs. Stay tuned to community updates and GitHub discussions for the latest trends and sampler innovations as AI art generation continues to evolve in 2025.

Call to Action:
If you found this guide helpful, share it on social media, leave a comment below with your experiences, and subscribe to our newsletter for more tips and updates on AI art tools!