This article covers how the zeroheight MCP retrieves content, and what you can do to make your styleguide work well with AI tools.
New to MCP? Read the zeroheight MCP overview first.
How the MCP retrieves content
When an AI agent uses the zeroheight MCP, it usually follows the following pattern:
- Fetches a list of available styleguides (skipped for Remote MCP via link, which is scoped to a single styleguide)
- Retrieves the sections, categories, and page list for the selected styleguide
- Fetches the content of pages it judges as relevant — based on your prompt or what it can infer from context
So the AI selects pages by their titles and structure, not by reading everything. It doesn't read the whole styleguide to decide what's relevant — it makes a judgement call from the navigation layer. This means how you structure your styleguide directly affects what the AI finds and uses.
Structuring your styleguide for AI
Use clear, descriptive page titles The AI navigates primarily by title.
One topic per page The AI picks pages, not sections within pages. Avoid combining unrelated content on a single page.
Keep navigation flat where possible Deeply nested structures are harder for an AI to traverse. Flatter hierarchies improve both discoverability and retrieval reliability.
Name sections and categories meaningfully Section and category names are surfaced during MCP navigation. "Foundations > Colour" gives the AI a solid signal.
Write content as rules, not descriptions Succinct, unambiguous language works better than detailed prose. State what to do, not just what exists.
Which AI tools work with the zeroheight MCP?
The MCP works with all the major AI coding, prototyping, and chat tools — including Claude, ChatGPT, Codex, Cursor, Copilot, Figma Make and others.
A few things to keep in mind:
- Model quality matters. Smaller or older models may struggle to navigate MCP responses reliably, especially when content is ambiguous or pages are large. Newer frontier models handle MCP tool use more consistently.
- OpenAI tools sometimes need prompting. ChatGPT and Codex occasionally need explicit instruction to use the MCP — but once they're calling the tools, they work well.
- There's no single "best" tool. The right choice depends on your team's existing workflow.