ConnectWise API Gateway MCP Server

This Model Context Protocol (MCP) server provides a comprehensive interface for interacting with the ConnectWise Manage API. It simplifies API discovery, execution, and management for both developers and AI assistants.

Core Capabilities

• API Discovery: Search for and explore ConnectWise API endpoints using keywords or natural language
• Simplified API Execution: Execute API calls with friendly parameter handling and automatic error management
• Fast Memory System: Save and retrieve frequently used API queries for more efficient workflows
• Raw API Access: Send custom API requests with complete control over endpoints, methods, and parameters

Key Features

• Database-Backed API Discovery: Uses a SQLite database built from the ConnectWise API definition JSON for fast, efficient endpoint lookups
• Natural Language Search: Find relevant API endpoints using conversational descriptions of what you need
• Categorized API Navigation: Browse API endpoints organized by functional categories
• Detailed Documentation Access: View comprehensive information about API endpoints including parameters, schemas, and response formats
• Adaptive Learning: The system learns which API calls are most valuable to you through usage tracking

Installation & Setup

Prerequisites

• Python 3.10 or higher
• Access to ConnectWise Manage API credentials
• ConnectWise API definition file (manage.json) - included in the repository

Installation Steps

Option 1: Using GitHub NPM Package (Recommended)

You can install the package directly from GitHub:

This method automatically handles all dependencies and provides a simpler configuration for Claude Desktop.

Option 2: Manual Installation

Windows

1. Clone or download the repository:
   git clone https://github.com/jasondsmith72/CWM-API-Gateway-MCP.git cd CWM-API-Gateway-MCP
2. Install the package:
   pip install -e .

macOS

For the NPM installation method, simply run:

For manual installation:

1. Install Python 3.10+ if not already installed:
   # Using Homebrew brew install python@3.10 # Or using pyenv brew install pyenv pyenv install 3.10.0 pyenv global 3.10.0
2. Clone the repository:
   git clone https://github.com/jasondsmith72/CWM-API-Gateway-MCP.git cd CWM-API-Gateway-MCP
3. Set up a virtual environment (recommended):
   python3 -m venv venv source venv/bin/activate
4. Install the package:
   pip install -e .

Linux (Ubuntu/Debian)

For the NPM installation method, simply run:

For manual installation:

1. Install Python 3.10+ if not already installed:
   # For Ubuntu 22.04+ sudo apt update sudo apt install python3.10 python3.10-venv python3.10-dev python3-pip # For older versions of Ubuntu/Debian sudo add-apt-repository ppa:deadsnakes/ppa sudo apt update sudo apt install python3.10 python3.10-venv python3.10-dev python3-pip
2. Clone the repository:
   git clone https://github.com/jasondsmith72/CWM-API-Gateway-MCP.git cd CWM-API-Gateway-MCP
3. Set up a virtual environment (recommended):
   python3.10 -m venv venv source venv/bin/activate
4. Install the package:
   pip install -e .

Post-Installation Steps

After installing on any platform (Windows, macOS, or Linux), complete the following steps:

1. (Optional) Build the API Database

This repository already includes a pre-built database, so this step is optional. Only run this if you need to use a newer ConnectWise API definition file:

This step only needs to be done once, or whenever the ConnectWise API definition changes.

2. Configure API Credentials

Set the following environment variables with your ConnectWise credentials:

These credentials are used in the authentication process as follows:

• CONNECTWISE_API_URL: The base URL for all API requests to your ConnectWise instance
  url = f"{API_URL}{endpoint}" # e.g., https://na.myconnectwise.net/v4_6_release/apis/3.0/service/tickets
• CONNECTWISE_COMPANY_ID: Included in the 'clientId' header of each request to identify your company
  headers = {'clientId': COMPANY_ID, ...}
• CONNECTWISE_PUBLIC_KEY and CONNECTWISE_PRIVATE_KEY: Used together with AUTH_PREFIX to create the basic authentication credentials
  username = f"{AUTH_PREFIX}{PUBLIC_KEY}" # e.g., "yourprefix+your_public_key" password = PRIVATE_KEY credentials = f"{username}:{password}" # Combined into "yourprefix+your_public_key:your_private_key"
• CONNECTWISE_AUTH_PREFIX: Required prefix added before your public key in the authentication username. ConnectWise API requires this prefix to identify the type of integration (e.g., "api+", "integration+", etc.)

The final HTTP headers sent with every request will look like:

Configuration for Claude Desktop

There are two methods to integrate with Claude Desktop:

Method 1: Using NPM Package (Recommended)

Install the package using NPM:

Then configure Claude Desktop (claude_desktop_config.json):

Method 2: Using Node.js Script (Alternate Method)

If you've cloned the repository and installed the dependencies, you can use the included Node.js script:

Method 3: Using Direct Python Script Path

If you prefer to use the Python script directly:

For macOS and Linux, use the appropriate path format:

The server can be run directly from the command line for testing:

Available Tools

The API Gateway MCP server provides several tools for working with the ConnectWise API:

API Discovery Tools

API Execution Tools

Fast Memory Tools

Usage Examples

Search for Ticket-Related Endpoints

Search Using Natural Language

Execute a GET Request

Create a New Service Ticket

Send a Raw API Request

View Fast Memory Contents

Save a Useful Query to Fast Memory

Understanding Fast Memory

The Fast Memory feature allows you to save and retrieve frequently used API queries, optimizing your workflow in several ways:

Benefits

• Time Savings: Quickly execute complex API calls without remembering exact endpoints or parameters
• Error Reduction: Reuse successful API calls to minimize potential errors
• Adaptive Learning: The system learns which API calls are most valuable to you
• Parameter Persistence: Parameters and request bodies are stored for future use

How It Works

1. Automatic Learning: When you execute a successful API call, you're prompted to save it to Fast Memory
2. Intelligent Retrieval: The next time you use the same API endpoint, the system checks Fast Memory first
3. Parameter Reuse: If you don't provide parameters for a call, the system automatically uses those saved in Fast Memory
4. Usage Tracking: The system tracks how often each query is used and prioritizes frequently used queries

Fast Memory Functionality

• Automatic Parameter Suggestion: The system will suggest parameters from Fast Memory if none are provided
• Usage Counter: Each time a query from Fast Memory is used, its usage count increases
• Search Capability: Search through your saved queries by description or endpoint path
• Prioritization: Queries are displayed in order of usage frequency, with most frequently used queries at the top

Managing Your Fast Memory

• View Saved Queries: list_fast_memory()
• Search Specific Queries: list_fast_memory("search term")
• Delete a Query: delete_from_fast_memory(query_id)
• Clear All Queries: clear_fast_memory()

Fast Memory Technical Details

The Fast Memory system is powered by a SQLite database (fast_memory_api.db) that stores:

• Query paths and methods
• Parameters and request bodies as JSON
• Usage metrics and timestamps
• User-friendly descriptions

The database structure includes:

• id: Unique identifier for each saved query
• description: User-provided description of what the query does
• path: API endpoint path
• method: HTTP method (GET, POST, PUT, etc.)
• params: Query parameters in JSON format
• data: Request body in JSON format
• timestamp: When the query was last used
• usage_count: How many times the query has been used

Troubleshooting

Common Issues

Database Not Found Error

Solution: Run the build_database.py script with the path to your ConnectWise API definition file:

API Authentication Issues

Solution: Check your environment variables to ensure all ConnectWise credentials are correct:

• Verify your CONNECTWISE_COMPANY_ID, CONNECTWISE_PUBLIC_KEY, and CONNECTWISE_PRIVATE_KEY
• Ensure the API key has the necessary permissions in ConnectWise
• Check that CONNECTWISE_AUTH_PREFIX is set correctly for your environment

Timeouts on API Calls

Solution:

• Check your internet connection
• The ConnectWise API may be experiencing high load
• For large data requests, consider adding more specific filters to your query

Logs and Diagnostics

Log Locations

• Main log file: api_gateway/api_gateway.log
• SQLite databases:
  • API Database: api_gateway/connectwise_api.db
  • Fast Memory Database: api_gateway/fast_memory_api.db

Testing the Database

Verify that the database is correctly built and accessible:

This will display statistics about the database and confirm it can be queried properly.

Advanced Usage

Optimizing API Queries

For better performance with the ConnectWise API:

1. Use Specific Conditions: Narrow your queries with precise conditions
   execute_api_call("/service/tickets", "GET", { "conditions": "status/name='Open' AND dateEntered > [2023-01-01T00:00:00Z]" })
2. Limit Field Selection: Request only the fields you need
   execute_api_call("/service/tickets", "GET", { "conditions": "status/name='Open'", "fields": "id,summary,status,priority" })
3. Paginate Large Results: Use page and pageSize parameters
   execute_api_call("/service/tickets", "GET", { "conditions": "status/name='Open'", "page": 1, "pageSize": 50 })

License

This software is proprietary and confidential. Unauthorized copying, distribution, or use is prohibited.

Acknowledgments

• Built using the Model Context Protocol (MCP) framework
• Powered by ConnectWise Manage API