Analysis updated 2026-05-18
Give an AI coding agent the current post WWDC25 Apple system colors instead of outdated ones.
Check generated SwiftUI interface code against Liquid Glass blur and contrast rules.
Look up which SwiftUI API corresponds to a specific Apple Human Interface Guidelines component.
Track which AI agents on a machine are actually calling the design guideline server.
| aka-kika/hig-mcp | 0xsero/tailscale-mullvad-toggle | afspies/clawdmoji | |
|---|---|---|---|
| Stars | 31 | 31 | 31 |
| Language | Python | Python | Python |
| Setup difficulty | easy | easy | easy |
| Complexity | 2/5 | 2/5 | 2/5 |
| Audience | developer | developer | general |
Figures from each repo's GitHub metadata at analysis time.
hig-mcp is a small server that speaks the Model Context Protocol, or MCP, a way for AI coding tools like Claude Code or Cursor to fetch outside information while they work. This particular server hands those tools the Apple Human Interface Guidelines, the design rules Apple publishes for building iOS and macOS apps, but as clean structured data instead of web pages full of prose. The problem it addresses is specific: Apple's guidelines change over time, most recently at WWDC25 when the system color palette was updated, so the blue color developers used to hardcode is no longer correct. AI models trained before that update still confidently suggest the old, wrong values, and reading the guidelines as web pages burns a lot of an agent's token budget while still leaving it to guess exact numbers. Apple's newer Liquid Glass visual style also has rules, like blur limits and contrast requirements, that are not written down in one clear place, so AI generated interface code tends to quietly violate them. hig-mcp solves this by keeping verified design values, such as colors, typography, spacing, and SwiftUI component mappings, bundled directly inside the server so they can be served instantly and offline. When the actual guideline text is needed rather than just the numbers, it fetches that live from a separate site called sosumi.ai so the wording stays current. Every value it returns is tagged with where it came from, whether it is an official Apple fact, an accessibility rule, or a community best practice, so a developer knows how much to trust it. Installation is a simple command line install followed by registering it with an MCP-compatible tool such as Claude Code. The project also logs every tool call it receives to a local file, so a developer can see which of their AI agents are actually using it. It currently covers iOS, iPadOS, and macOS, is written in Python, and its code and bundled data are released under the MIT license, though it is not affiliated with or endorsed by Apple.
An MCP server that gives AI coding tools verified Apple design guideline data, like current colors and SwiftUI rules, instead of stale web text.
Mainly Python. The stack also includes Python, FastMCP, MCP.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.