Generating high-quality game assets can be time-consuming and challenging. However, with recent advances in AI image generation models like Stable Diffusion, it is now possible to quickly create game assets by providing text prompts.
In this article, I will provide a comprehensive guide on writing effective Stable Diffusion prompts to generate 2D and 3D game assets. I will cover prompt structure, formatting, tips and tricks, as well as provide 15 example prompts across various game asset categories.
Prompt Structure
When writing Stable Diffusion prompts, it is important to structure them properly so the AI understands exactly what you want to generate. Here is a basic prompt structure to follow:
[Subject], [Style], [Viewpoint/Framing], [Medium/Rendering], [Extra Details]
For example:
Cyberpunk character portrait, trending on artstation, digital painting
Let’s break down each component:
- Subject: The main focus of the image, like “character”, “environment”, “vehicle”, etc.
- Style: Artistic style you want like “cartoon”, “pixel art”, “low poly”. Reference a specific game or artist for consistency.
- Viewpoint/Framing: Composition of the image, like “side view”, “close up shot”, “aerial view”.
- Medium/Rendering: The artistic medium or rendering style, like “digital painting”, “3D render”, “vector art”.
- Extra Details: Any additional descriptive details about lighting, colors, props, etc.
Formatting Tips
Here are some key formatting tips:
- Use double parentheses around terms you want the AI to ignore like ((this))
- Separate clauses with commas
- Use present tense action verbs like “a man stands in a forest”
- Capitalize proper nouns only
- Use single spaces between words/clauses
Advanced Prompting Tips
To further improve results, leverage these advanced prompting techniques:
- Weighted Sampling: Use a colon and number to weight terms, like “nature:2” to bias generation towards that term
- Image Inserts: Insert an example image with
<image:INSERT_URL_HERE>
to guide the AI - Negative Prompts: Add
--negativeprompt
flags to exclude unwanted elements from generation like--negativeprompt:poorly drawn
15 Stable Diffusion Prompts for Game Assets
Below are 15 prompt examples covering characters, environments, props, textures, and UI elements.
Characters
Cyberpunk protagonist character full body view, trending on artstation HQ
armored orc warrior wielding a battleaxe, dynamic pose, unreal engine digital rendering
Cartoon mushroom character facial portrait, soft lighting, Nintendo style
Environments
Japanese zen garden environment, aerial view, watercolor painting
mystical ruins environment with vines and vegetation, isometric view, digital matte painting
Futuristic city streets, night time, cinematic lighting, Unity render
Props
mythical sword and shield set, detailed craftsmanship, epic lighting, digital painting
Potions and elixirs sprite sheet, 16 icons, top down view, flat vector illustration
Textures
Seamless futuristic metal panel texture, scratched paint, 4k resolution
stone floor texture with moss and cracks, procedural material, substance designer
Stylized water texture, cartoon waves animation, looping gif
UI/Icons
Inventory icon set sprite sheet, 16 icons, isometric view, vector illustration
futuristic hologram buttons and UI elements, cinematic lighting, detailed panels
Gothic frame decoration icons, black and gold baroque style, vector image
Conclusion
I hope this guide gives you a solid foundation for writing effective prompts to generate 2D and 3D game assets with Stable Diffusion AI. Properly structuring prompts, using formatting best practices, and leveraging advanced techniques will produce higher quality, consistent results.
Feel free to take inspiration from the 15 prompt examples provided across characters, environments, props, textures and UI elements. Don’t be afraid to continually iterate on prompts and guide the AI by inserting example images.
With some practice, you’ll be able to rapidly create game assets to supplement your development workflow. Stable Diffusion eliminates much of the heavy lifting involved in asset creation, allowing you to focus creativity on designing gameplay systems and mechanics.
Let me know in the comments if you have any other questions!