Adonis Documentation Server

MCP (Model Context Protocol) server to access adonis-mcp documentation.

📋 Resources

file:///{name}.md - Documentation Markdown File

Resource template to retrieve markdown documentation files from the GitHub repository.

Available files:

• inspector - MCP Inspector tool documentation
• installation - Installation guide
• introduction - Introduction to MCP
• prompts - Prompts documentation
• resources - Resources documentation
• sessions - Session management
• tools - Tools documentation
• unit-tests - Unit testing

Usage example:

file:///resources.md
file:///tools.md

🛠️ Tools

1. list_documentation

Lists all available documentation files with their descriptions.

Parameters: None

Returns: Structured list of all available documentation files

Annotations:

• @isReadOnly() - Read-only
• @isIdempotent() - Can be called multiple times without side effects

2. search_in_docs

Search for a keyword or phrase across all documentation files.

Parameters:

• query (string, required) - The keyword or phrase to search for
• caseSensitive (boolean, optional) - Case-sensitive search (default: false)

Returns: List of matches with context (2 lines before/after)

Annotations:

• @isReadOnly() - Read-only
• @isOpenWorld() - Accesses external resources (GitHub)
• @isIdempotent() - Can be called multiple times without side effects

Example:

{
  "query": "Resource templates",
  "caseSensitive": false
}

3. extract_code_examples

Extracts all code blocks from a documentation file with their languages and line numbers.

Parameters:

• filename (string, required) - File name (without .md extension)

Returns: Structured list of all code blocks with their language, content, and line number

Annotations:

• @isReadOnly() - Read-only
• @isOpenWorld() - Accesses external resources (GitHub)
• @isIdempotent() - Can be called multiple times without side effects

Example:

{
  "filename": "resources"
}

🎯 Prompts

explain_feature

Guide to get a detailed explanation of an adonis-mcp feature with code examples.

Parameters:

• feature (string, required) - The feature to explain
• level (string, optional) - Audience level: "beginner", "intermediate", "advanced" (default: "intermediate")

Suggested features:

• resources
• tools
• prompts
• resource templates
• tool annotations
• completions
• authentication
• authorization
• sessions
• unit testing
• inspector
• middleware

Example:

{
  "feature": "resource templates",
  "level": "intermediate"
}

🚀 Usage

Installation

Install globally via npm:

npm install -g adonis-mcp-docs

Or use with npx (no installation required):

npx adonis-mcp-docs

Testing locally before publishing

1. Build the project:

npm run build

2. Test with npx locally:

cd build
node bin/mcp.js

Start the MCP server

adonis-mcp-docs

Or with npx:

npx adonis-mcp-docs

Configuration for MCP Clients

Add to your Claude Desktop or Cursor MCP configuration:

{
  "mcpServers": {
    "adonis-mcp-docs": {
      "command": "npx",
      "args": ["adonis-mcp-docs"]
    }
  }
}

📦 Services

DocumentationService

Manages retrieval of remote markdown files from GitHub.

Methods:

• fetchMarkdownFile(name: string) - Fetches a markdown file
• fileExists(name: string) - Checks if a file exists
• getFileSize(name: string) - Gets the size of a file

ResourceCompletionService

Provides completion suggestions for documentation file names.

Methods:

• getCompletions(prefix?: string) - Returns files matching the prefix
• getAllNames() - Returns all available file names

🔍 Combined Usage Examples

Search then read

1. Use search_in_docs to find where a concept is documented
2. Use the file:///{name}.md resource to read the complete content

Extract and analyze

1. Use extract_code_examples to get all examples
2. Analyze the code with a custom prompt

Explore then explain

1. Use list_documentation to see what's available
2. Use the explain_feature prompt to get a detailed explanation

📝 Technical Notes

• Base URL: https://raw.githubusercontent.com/batosai/adonis-mcp/main/docs
• Format: Markdown (.md)
• Search: Case-insensitive by default with 2 lines of context
• Parsing: Code block extraction with automatic language detection

🎨 Advanced Features

• ✅ Completions enabled for all templates
• ✅ Tool annotations to guide AI
• ✅ Robust error handling
• ✅ Search with context
• ✅ Automatic extraction of all code blocks with language detection