OpenSearch MCP Server

A Model Context Protocol (MCP) server that provides tools for interacting with OpenSearch clusters through Server-Sent Events (SSE). This server enables seamless integration of OpenSearch operations into AI workflows.

Features

• OpenSearch Integration: Secure connection and interaction with OpenSearch clusters
• SSE Communication: Real-time, event-driven communication using Server-Sent Events
• Tool Suite:
  • List indices in the cluster
  • More tools coming soon...

Prerequisites

• Python 3.12 or higher
• OpenSearch cluster (local or remote)
• uv for dependency management (recommended)

Installation

# Install uv (if not already installed)
pip install uv

# Clone the repository
git clone https://github.com/your-org/opensearch-mcp-sse.git
cd opensearch-mcp-sse

# Create and activate virtual environment
uv venv .venv
source .venv/bin/activate

# Install dependencies using the lock file (recommended for production)
uv pip install -r uv.lock

# Or install in editable mode for development
uv pip install -e .

Dependency Management

The project uses uv for dependency management with a uv.lock file to ensure reproducible builds. The lock file contains exact versions of all dependencies and their sub-dependencies.

• pyproject.toml: Defines project metadata and direct dependencies
• uv.lock: Contains exact versions of all dependencies (including transitive dependencies)

To update dependencies:

# Update dependencies and regenerate lock file
uv pip compile pyproject.toml -o uv.lock

Configuration

OpenSearch Connection

Configure the OpenSearch connection using environment variables:

# Required
export OPENSEARCH_HOST=localhost      # OpenSearch host
export OPENSEARCH_PORT=9200          # OpenSearch port

# Optional (with defaults)
export OPENSEARCH_USERNAME=admin     # Username for authentication
export OPENSEARCH_PASSWORD=admin     # Password for authentication
export OPENSEARCH_USE_SSL=false      # Enable/disable SSL
export OPENSEARCH_VERIFY_CERTS=false # Verify SSL certificates
export OPENSEARCH_CA_CERTS=          # Path to CA certificates

Server Configuration

The server can be configured through command-line arguments:

To run the server go to src Folder and run:

python -m server --host 0.0.0.0 --port 9900

Default values:

• Host: 0.0.0.0 (all interfaces)
• Port: 9900
• API Key: secret-token

Usage

Starting the Server

# Start with default configuration
python -m server

# Start with custom host and port
python -m server --host 127.0.0.1 --port 8080

Authentication

All requests must include an API key in the Authorization header:

Authorization: Bearer secret-token

Available Tools

1. List Indices (list_indices)
   • Lists all indices in the OpenSearch cluster
   • Returns: List of index names
   • Example response: ["index1", "index2", ...]

More coming soon

Development

Project Structure

opensearch-mcp-sse/
├── src/
│   ├── mcp_server.py    # MCP initialization
│   ├── server.py        # Server implementation with SSE and authentication
│   └── opensearch/      # OpenSearch integration
│       ├── config.py    # Configuration management
│       ├── client.py    # OpenSearch client with core operations
│       └── tools/       # MCP tools
│           ├── list_indices.py
│           └── __init__.py
└── scripts/            # Utility scripts

Adding New Tools

1. Create a new tool file in src/opensearch_mcp_sse/opensearch/tools/
2. Import and register the tool in src/opensearch_mcp_sse/server.py

Running Scripts

# Run scripts
python scripts/test_connection.py