MCP Docs Tools
About
this is a MCP server that create uml text based class diagrams, project tree structure and a list of modules anf functions in the project to assist cursor in understanding the project.
Explore
- Three focused tools: UML class diagrams, directory trees, and module function docs
- Zero configuration – works out of the box with sensible defaults
- Smart exclusions automatically filter cache, build, and IDE files
- Captures type hints, decorators, docstrings, and inheritance relationships
- Hybrid Node.js + Python architecture for reliable AST parsing
- Seamless integration with AI assistants via MCP protocol
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
MCP Docs ToolsCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
- Node.js ≥18.0.0
- Python 3.x (for AST parsing)
- Git (for cloning the repository)
npm install
npm install -g .
npm install -g git+https://github.com/your-username/mcp-docs-tools.git
mcp-docs-tools
{
"mcpServers": {
"docs-tools": {
"command": "mcp-docs-tools"
}
}
}
Create this .cursorrules file in your Python projects:
1. Clone and install (recommended):
git clone https://github.com/your-username/mcp-docs-tools.git
cd mcp-docs-tools
npm install
2. Find your installation path:
pwd
3. Update your MCP config with the path from step 2
4. Test the tools in your Python projects!
create_class_diagram
Generate UML class diagrams
create_tree_structure
Generate directory tree
create_module_functions
Document module functions
A Node.js-based MCP (Model Context Protocol) tool server that provides three focused documentation generation tools for Python projects. Built as an npm package that can be installed directly from the Git repository.
1. create_class_diagram - Analyzes Python files and generates UML class diagrams in PlantUML format
2. create_tree_structure - Creates clean directory tree structure documentation with smart exclusions
3. create_module_functions - Documents module-level functions with signatures, decorators, and type hints
- Version control: .git, .svn, .hg
- IDEs: .idea, .vscode, .cursor
- Testing: .pytest_cache, .coverage, .tox
- create_class_diagram - Generate UML class diagrams
- create_tree_structure - Generate directory tree
- create_module_functions - Document module functions
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"mcp docs tools": {
"project-docs": {
"command": "node",
"args": [
"bin/server.js"
]
}
}
}
}McpServers
{
"project-docs": {
"command": "node",
"args": [
"bin/server.js"
]
}
}A Node.js-based MCP (Model Context Protocol) tool server that provides three focused documentation generation tools for Python projects. Built as an npm package that can be installed directly from the Git repository.
Features
🔧 Three Powerful Tools
1. create_class_diagram - Analyzes Python files and generates UML class diagrams in PlantUML format
2. create_tree_structure - Creates clean directory tree structure documentation with smart exclusions
3. create_module_functions - Documents module-level functions with signatures, decorators, and type hints
🚀 Key Benefits
- Zero Configuration: Works out of the box with sensible defaults
- Smart Exclusions: Automatically filters out cache, build, and IDE files
- Rich Documentation: Captures type hints, decorators, docstrings, and inheritance
- MCP Integration: Seamlessly integrates with AI assistants via Model Context Protocol
- Hybrid Architecture: Node.js orchestration + Python AST parsing for reliability
Installation
From Git Repository (Recommended)
```bash
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
