building-game-images-with-localai

Generates 2D game art assets with LocalAI image generation and ImageMagick CLI post-processing.

1|Updated Mar 19, 2026
One-click install
npx skills add https://github.com/fiando/bake-tycoon --skill building-game-images-with-localai-fiando
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: building-game-images-with-localai
Source: https://github.com/fiando/bake-tycoon/tree/main/.github/skills/building-game-images-with-localai
Command: npx skills add https://github.com/fiando/bake-tycoon --skill building-game-images-with-localai-fiando

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Producing consistent 2D game art (sprites, tiles, icons, backgrounds) locally is hard when prompts vary wildly and raw model output is not production-ready. This Skill provides a repeatable LocalAI-first workflow with stable prompt templates and CLI cleanup steps. ## Core Features & Use Cases - LocalAI Image Generation: Uses the POST /v1/images/generations endpoint with diffusers or stablediffusion-ggml backends for fully local asset creation. - Prompt Templates: A fixed prompt pattern (subject, view, style, detail, readability goal, negative prompt) keeps style consistent across batches of sprites, tiles, UI icons, and backgrounds. - CLI Post-Processing: ImageMagick commands for nearest-neighbor resizing, transparent-border trimming, and spritesheet packing via montage. - Use Case: Generate a set of bakery-themed UI icons for an idle game by reusing one approved prompt structure, then trim and resize each icon to 128x128 with magick before importing into the game engine. ## Quick Start Ask the assistant to generate a 512x512 game icon with LocalAI using the standard prompt pattern, then trim and resize it with ImageMagick.

Frequently Asked Questions about building-game-images-with-localai

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate game sprites with LocalAI?▼

Send a POST request to the LocalAI /v1/images/generations endpoint with a structured prompt describing subject, view, style, and readability goal plus a negative prompt. Then crop and scale the output with ImageMagick to get a usable sprite.

What prompt structure works best for 2D game art generation?▼

Use the pattern: subject, camera/view, style, material detail, readability goal, followed by a negative prompt. Keep one model family and one prompt shape per batch so all assets share a consistent visual style.

Which LocalAI backend should I use for image generation?▼

LocalAI documents two image paths: diffusers for standard Hugging Face diffusion workflows, and stablediffusion-ggml for stable-diffusion.cpp style local setups. Both are called through the same /v1/images/generations API.

How do I resize pixel art without blurring it?▼

Use ImageMagick with the point filter: magick input.png -filter point -resize 128x128 output.png. The point filter applies nearest-neighbor scaling, which preserves hard pixel edges instead of smoothing them.

When should I not use LocalAI for game asset creation?▼

Avoid this workflow for 3D modeling, rigging, or skeletal animation, since diffusion image generation only produces 2D images. It is also unnecessary if you only need engine import settings rather than new source art.