design-please
design-please helps non-designers say what they want, clearly enough that a tool, a designer, or a developer can build it.
What it does
Most non-designers know good design when they see it but cannot describe what they want in terms anyone can act on. So AI design tools fall back on their built-in house style, whatever the product in front of them actually needs.
design-please works on the input instead of the tool. It turns reference products, brand adjectives, screenshots, and existing work into a clear brief: structured documents that specify three things precisely enough to build from.
- Look: the visual surface, color, type, spacing.
- Feel: the emotional register, what the product should make someone feel.
- Act: how it behaves, the interaction, the motion, the density.
How it works
design-please works in three phases:
- Discover: it interviews you for references and anti-references across Look, Feel, and Act, and reviews anything you have already made. When you genuinely have no opinion on something, it offers to choose for you and explains why.
- Author: it writes the Design Brief, the central document, plus a design system spec at whatever level of detail you need, from a few colors and fonts up to a full system.
- Encode: it turns the brief into a paste-ready configuration for Claude Design, including the part that keeps the tool's default style from creeping back in.
Like its siblings, design-please is clear about its limits. You can run it yourself and get something usable. But the move from "it should feel premium" to a genuinely distinctive feel usually takes a person with taste, someone who can bring references from outside what you have already seen.
Who it's for
- Founders and product leads about to use an AI design tool who do not know how to prompt it
- People shipping product who know what "wrong" looks like but cannot describe "right"
- Teams who keep getting off-brief design output and cannot tell why
Using it without writing code
You do not need to be a developer to use design-please, and you do not need to be comfortable with GitHub. The framework is a set of plain-language instructions that an AI assistant reads and works from. Your part is the conversation.
The quickest way in is the "Open in Claude" button on this page. It hands Claude a short brief on design-please and leaves room for you to describe your own situation, so you can talk through the fit before committing to anything.
To run the framework itself, you open the repository with Claude Code, Anthropic's assistant that works directly with files, and tell it what you need in ordinary sentences, like "I am about to open Claude Design, get me ready." It handles setup the first time and walks you through the rest. If any of that feels unfamiliar, it is a good thing to bring to a first conversation.
See it on GitHub
The full framework, its step-by-step skills, and worked examples are all in the open repository: