MCP MarkLogic Server

This is a Model Context Protocol (MCP) server implementation for MarkLogic, allowing you to interact with a MarkLogic database through MCP tools.

2. The server provides the following operations:

Create Document

• Create documents in MarkLogic with optional collections

• Read documents by URI

• Delete documents by URI

• Search documents with query strings and optional collection filters

Installation

Related MCP server: mcp-graphql

Configuration

The server requires the following environment variables:

Usage

Once installed and configured, you can start the server:

Available Tools

1. create-document

   • Create a new document in MarkLogic

   • Parameters:

     • uri: Document URI (required)

     • content: Document content as JSON (required)

     • collections: List of collections to add the document to (optional)

2. read-document

   • Read a document from MarkLogic

   • Parameters:

     • uri: Document URI (required)

3. delete-document

   • Delete a document from MarkLogic

   • Parameters:

     • uri: Document URI (required)

4. search-documents

   • Search for documents in MarkLogic

   • Parameters:

     • query: Search query string (required)

     • collections: List of collections to search in (optional)

Development

Setup

1. Clone the repository

2. Create a virtual environment and activate it

3. Install dependencies:

   pip install -e .

Running Tests

MarkLogic Setup

1. Ensure you have MarkLogic Server installed and running

2. Create a REST API instance on port 8000 if not already present

3. Create a user with appropriate permissions:

   • Go to Admin Interface (usually http://localhost:8001)

   • Create role python-docs-role with:

     • Roles: rest-extension-user, rest-reader, rest-writer

     • Privileges: xdbc:eval, xdbc:invoke, xdmp:eval-in

   • Create user python-user with:

     • Password: pyth0n

     • Role: python-docs-role

License

MIT