AWS Amplify Documentation MCP Server

A Model Context Protocol (MCP) server that provides powerful search functionality for AWS Amplify documentation. This server clones the official AWS Amplify documentation repository and makes it searchable through a simple MCP tool interface.

Features

• Powerful Search: Search AWS Amplify documentation using natural language queries

• Advanced Search Syntax: Support for boolean operators, wildcards, field-specific search, and more

• Smart Results Ranking: Intelligent ranking of search results based on relevance to query

• Pagination: Navigate through large result sets with pagination support

• Performance Caching: Caching of search results for improved performance

• Auto-Updates: Automatic updates of documentation from the official AWS Amplify repository

• Generation Selection: Choose between Gen 1, Gen 2, or both documentation sets to optimize disk usage and search radius.

• TypeScript Implementation: Built with TypeScript for better type safety and developer experience

Disclaimer: This is a personal project and is not affiliated with, endorsed by, or officially connected to AWS Amplify or Amazon Web Services. This tool is provided as-is without any guarantees or warranty.

Installation

1. Clone this repository:

git clone https://github.com/ykethan/amplify-doc-mcp.git
cd amplify-doc-mcp

1. Install dependencies:

npm install

1. Build the TypeScript code:

npm run build

Configuration

The server is configured using the docs-mcp.config.json file:

{
  "gitUrl": "https://github.com/aws-amplify/docs.git",
  "gitRef": "main",
  "autoUpdateInterval": 60,
  "toolName": "search_amplify_docs",
  "toolDescription": "Search AWS Amplify documentation using the probe search engine.",
  "ignorePatterns": [
    "node_modules",
    ".git",
    "dist",
    "build",
    "coverage",
    ".vitepress/cache",
    "*.jpg",
    "*.jpeg",
    "*.png",
    "*.gif",
    "*.svg",
    "*.mp4",
    "*.webm"
  ],
  "amplifyGeneration": "gen2"
}

Configuration Options

Auto-Update Mechanism

The server includes an automatic update mechanism that keeps the documentation up-to-date:

1. When the server starts, it clones the documentation repository specified in gitUrl.

2. If autoUpdateInterval is set to a value greater than 0, the server will periodically check for updates.

3. Every autoUpdateInterval minutes, the server:

   • Fetches the latest changes from the remote repository

   • Checks if the local branch is behind the remote branch

   • If updates are available, pulls the changes automatically

   • If no updates are needed, continues with the current documentation

This ensures that your documentation search results always include the latest information without requiring a server restart.

Usage

Starting the Server

npm start

Or use the provided start script with options:

./start-server.sh [--gen <1|2|both>] [--rebuild]

Options

• --gen 1: Include only Gen 1 documentation (reduces disk space usage)

• --gen 2: Include only Gen 2 documentation (default, reduces disk space usage)

• --gen both: Include both Gen 1 and Gen 2 documentation

• --rebuild: Force rebuild of data directory

Examples

# Start with Gen 2 documentation (default)
./start-server.sh

# Start with only Gen 1 documentation
./start-server.sh --gen 1

# Start with both Gen 1 and Gen 2 documentation
./start-server.sh --gen both

# Start with only Gen 2 documentation and force rebuild
./start-server.sh --gen 2 --rebuild

Important: When switching between different generation options (e.g., from Gen 1 to Gen 2 or vice versa), it's recommended to use the --rebuild flag to ensure a clean repository. This will remove the data directory and clone the repository again, ensuring that the correct files are included based on the selected generation.

For example:

# Switch from Gen 2 to Gen 1 with a clean rebuild
./start-server.sh --gen 1 --rebuild

# Switch from Gen 1 to Gen 2 with a clean rebuild
./start-server.sh --gen 2 --rebuild

MCP Tool: search_amplify_docs

The server provides a tool called search_amplify_docs that can be used to search the Amplify documentation.

Parameters

Example Request

{
  "query": "authentication react",
  "page": 1,
  "includeContent": true,
  "maxResults": 15
}

Advanced Search Syntax

The search tool supports advanced search syntax:

• Exact phrases: "authentication flow"

• Exclude terms: authentication -flutter

• Field-specific search: title:authentication

• Wildcards: auth*

• Boolean operators: authentication AND (react OR javascript) NOT flutter

How the Query Process Works

When you submit a search query, the server processes it through several steps:

1. Query Processing: The server parses your query to understand advanced syntax like boolean operators and wildcards.

2. Directory Optimization: The server uses the directory structure to narrow down which files to search based on your query terms and the selected Amplify generation (Gen 1, Gen 2, or both).

3. Smart Ranking: Results are ranked using a sophisticated algorithm that considers:

   • Whether the query mentions specific generations (Gen 1 or Gen 2)

   • If the query is about setup, CLI commands, or resource creation

   • The relevance of the document to the query context

   • Exact matches in document titles

   • Match count and document importance

4. Content Extraction: For each matching file, relevant content is extracted and formatted for display.

5. Caching: Search results are cached to improve performance for repeated queries.

This intelligent processing ensures that the most relevant documentation appears at the top of your search results, saving you time and effort.

Generation Selection

The server supports three modes for Amplify documentation generation:

1. Gen 2 Only (Default)

"amplifyGeneration": "gen2"

• Pros: Focused on modern Amplify implementation, reduced disk space usage

• Recommended for: New projects using Amplify Gen 2 features

• Default: This is the default setting as Gen 2 is the current version of Amplify

2. Gen 1 Only

"amplifyGeneration": "gen1"

• Pros: Focused on classic Amplify implementation, reduced disk space usage

• Recommended for: Legacy projects specifically using Amplify Gen 1 features

3. Both Generations

"amplifyGeneration": "both"

• Pros: Complete documentation coverage

• Cons: Larger search area, can cause confusion as Gen 1 and Gen 2 have similar categories

• Use when: You need to reference both generations simultaneously

Project Structure

• src/index.ts: Main server implementation

• src/config.ts: Configuration loading and processing

• src/git.ts: Git repository management

• src/cache.ts: Search result caching

• src/directory.ts: Directory structure management

• src/types/: TypeScript type definitions

• scripts/build.js: Build script for preparing documentation

• bin/mcp: Executable script for running the server

Recommendations for Usage

1. Optimize for Your Environment:

   • For best search results, use "amplifyGeneration" as "gen1" or "gen2" only

2. Search Optimization:

   • Use specific technical terms rather than general phrases

   • Include category names to narrow results (e.g., "storage owner access" instead of just "access")

   • Use quotes for exact phrase matching

   • Include abbreviations and alternative terms to improve results

3. Performance Considerations:

   • Set an appropriate autoUpdateInterval based on your needs (higher values reduce server load)

   • Use the caching system for frequently accessed queries

   • Consider using filesOnly: true for initial broad searches to improve performance

Contributing and Feedback

We welcome contributions and feedback to improve this MCP server. If you have suggestions for:

• Improving search query results

• Enhancing the ranking algorithm

• Adding new features or parameters

• Optimizing performance

Please open an issue or submit a pull request on GitHub. Your feedback helps make this tool more effective for everyone. Along the way learn something new.

System Requirements

• Node.js 18.x or higher (tested with Node.js 20.18.2)

• npm 8.x or higher (tested with npm 10.8.2)

License

This project is licensed under the MIT License.