MidJourney's unofficial visual dictionary

What it does

This repository is a sprawling, community-curated reference that catalogs how specific keywords, styles, and parameters affect MidJourney image generation. It breaks down visual concepts—lighting, materials, artist styles, camera angles, even emojis—into categorized markdown pages packed with example outputs. The project also compares MidJourney’s built-in parameters like --stylize, --chaos, and image weights side-by-side.

The interesting bit

Rather than offering vague prompting tips, the project treats MidJourney like a black-box renderer and empirically documents its behavior. You can see exactly how changing an adjective, suffix, or resolution alters output, turning prompt engineering from guesswork into something closer to a science.

Key highlights

• Covers granular visual categories: colors, palettes, drawing mediums, SFX, shaders, dimensionality, geography, and more
• Includes comparison pages for prompt formatting, keyword weights, compound word formats, and adjective choices
• Documents parameter interactions such as quality vs. stylize, chaos vs. quality, and version-specific remastering behavior
• Provides resolution and upscaling reference charts across different aspect ratios
• Explicitly unofficial, with community submission guidelines and an active Discord thread

Caveats

• Pure documentation: no code, API wrapper, or automation tools—just Markdown and images
• The README is a dense navigation hub of linked buttons; finding specifics requires drilling down into sub-pages
• Currency depends on community contributions as MidJourney releases new versions

Verdict

A must-bookmark for MidJourney users who want to stop guessing and start knowing. Skip it if you don’t use MidJourney or need a programmable SDK.