MCP Documentation Service
local-only server
The server can only run on the client’s local machine because it depends on local resources.
MCP Documentation Service
The MCP Documentation Service is a custom implementation of the Model Context Protocol that allows AI assistants to interact with documentation. This service enables the creation, reading, updating, and deletion of documentation files, as well as searching and analyzing the documentation.
Features
- Document Management: Create, read, update, and delete markdown documentation files
- Metadata Management: Work with document frontmatter (YAML metadata)
- Search: Search through documentation using keywords and filters
- Analytics: Analyze documentation health and get suggestions for improvement
- Custom Directory Support: Specify a custom docs directory and create it if it doesn't exist
Installation
Via npx (Recommended)
The easiest way to use MCP Documentation Service is through npx:
Global Installation
You can also install it globally:
Local Installation
If you prefer to install it locally in your project:
Command-Line Options
The MCP Documentation Service supports the following command-line options:
--docs-dir <path>
: Specify the docs directory (default: ./docs)--create-dir
: Create the docs directory if it doesn't exist--help
,-h
: Show help message
Integration
With Cursor IDE
Add this to your .cursor/mcp.json
file:
To specify a custom docs directory:
With Claude Desktop
Add this to your claude_desktop_config.json
file:
To specify a custom docs directory:
Programmatic Usage
You can also use MCP Documentation Service programmatically in your Node.js application:
You can also use the query function directly:
Query Format
The service uses a SQL-like query format to execute commands:
For example:
Available Commands
Document Operations
- list_files(directory=""): List all markdown files (optionally in a specific directory)
- list_directories(directory=""): List all directories (optionally in a specific directory)
- get_document(path="path/to/doc.md"): Get a document's content and metadata
- create_document(path="path/to/doc.md", content="content", metadata={...}): Create a new document
- update_document(path="path/to/doc.md", content="updated content", metadata={...}): Update an existing document
- delete_document(path="path/to/doc.md"): Delete a document
Search & Analysis
- search_documents(query="search term", directory="", tags=["tag1"], status="published"): Search documents
- analyze_docs(directory=""): Analyze documentation health
- get_health_score(): Get overall documentation health score
- get_suggestions(): Get suggestions for improving documentation
License
MIT
This server cannot be installed
A Model Context Protocol implementation that enables AI assistants to interact with markdown documentation files, providing capabilities for document management, metadata handling, search, and documentation health analysis.