Quickstart

This guide takes about five minutes and needs no signup. You’ll pick a style from the public gallery, copy it compiled for your tool, and generate something in that style.

By the end you’ll have generated output that follows a shared, explicit style — and you’ll know the two ways to go further.

Before you begin

You need one AI tool you already use — ChatGPT, Claude, Gemini, Midjourney, or FLUX. Nothing to install.

Steps

Pick a style

Open the gallery and choose any public style. Every gallery style works with any AI tool. Copy its page URL, or just the slug at the end of it — that’s the style reference you’ll pass to any StyleRef surface.

Copy it compiled for your tool

On the style’s page, choose the format that matches where you’re headed, then copy the compiled spec:

ChatGPT, Claude, Gemini → default
Midjourney → midjourney
FLUX, DALL·E → flux
Stable Diffusion → diffusion

The copied text is a self-contained style block. Don’t paraphrase or trim it — it’s written to be applied verbatim.

Paste it into your tool

Place the style block first, then your request:

Chat tools (ChatGPT, Claude, Gemini): paste the block, then add your request below it (or set the block as a custom instruction / system prompt).
Midjourney: put your subject first, then the style keywords, and keep any --parameters at the very end.
Image models (FLUX, DALL·E, Stable Diffusion): put your subject on the first line, then the style block.

Generate

Run it. The tool now applies the style’s colors, mood, and composition to your request. Generate a few times — because every generation gets the same explicit constraints, the tool holds the style consistently across them.

Different tools render the same style differently — that’s expected. StyleRef makes your intent explicit within each tool; it doesn’t make two models produce the same image.

Where to go next

You’ve used someone else’s style. The two next steps fork here:

Make it yours (the product). Create a style from scratch, or extract one from text, a reference, an image, or a document.
Wire it into an agent. Connect the MCP server so an agent can search, fetch, and apply styles at runtime — or use the REST API and CLI.

Next steps

Core concepts — the terms used throughout these docs.
Use your style in AI tools — per-tool placement guides.
Gallery, sharing & remixing — save, publish, and remix styles.