MCP connection

Checking service status…

Search, read, and cite this site’s docs in your AI tools.

Public docs · Read-onlyStreamable HTTP

Available capabilities5Search, read, and discover

search_documents: Combines full-text and multilingual semantic retrieval with compact plain-text excerpts.
list_documents: Browse by language, area, and tag with stable cursors.
read_document: Read Markdown, metadata, and the full outline.
read_document_section: Read a bounded section by heading to control context size.
related_documents: Find related documents through shared tags.

Tool descriptions clearly mark public, read-only, and untrusted-content boundaries.

Connect your client

Choose your tool to get the right setup instructions.

Merge this configuration into .cursor/mcp.json in your project.

.cursor/mcp.jsonJSON

{
  "mcpServers": {
    "liz6-docs": {
      "url": "https://doc.liz6.com/mcp"
    }
  }
}

Save the file, then enable liz6-docs in Cursor’s MCP settings.

Another client? Use the server URL

https://doc.liz6.com/mcp

Connection optionsHTTP · WebMCP · Markdown

MCP Checking: Five read-only tools, document Resources, three workflow Prompts, and an MCP App for browsing results.
HTTP / OpenAPI ↗Checking: For scripts, workflow platforms, and Agents without MCP; supports lexical, semantic, and hybrid retrieval.
WebMCP ↗Checking: Compatible browsers discover the tools registered by each page; ordinary browsers remain unchanged.
Markdown / llms.txt ↗: Source text preserves structure and metadata for lightweight fetching, citation, and long-context reading.

If your client does not support MCP, use OpenAPI, llms.txt, or append .md to any document URL.

How browser Agents use itWebMCP

On any page, a WebMCP-capable Chrome Agent can see all five tools directly—no URL copying required.

WebMCP is progressively enabled through the Chromium Origin Trial; the MCP URL remains the more stable choice for general external Agents.