elasticsearch7

This MCP server provides a protocol wrapper around Elasticsearch 7.x, exposing common Elasticsearch operations through the MCP interface. It supports basic health and metadata endpoints (es-ping, es-info) as well as full text search capabilities via es-search, including support for aggregations, highlighting, sorting, and source filtering. Clients can connect using any MCP client and issue commands like es-ping to verify connectivity, es-info to fetch cluster details, and es-search to perform queries against specified indices. The server is designed for Elasticsearch 7.x clusters and aims to make Elasticsearch functionality accessible through MCP without requiring direct Elasticsearch client libraries.

To get started, configure the required environment variables (ELASTIC_HOST, ELASTIC_USERNAME, ELASTIC_PASSWORD) and run the MCP server. Then, connect with an MCP client to localhost on the configured port (default 9999) and invoke methods such as es-ping, es-info, or es-search. The README provides example client usage demonstrating a typical flow for performing a search or a more advanced query with aggregations and highlighting.

Prerequisites:

• Python 3.10+ installed on the host
• Access to an Elasticsearch 7.x cluster (7.17.x recommended)
• Python package manager (pip)

Installation steps:

1. Install Python dependencies in your environment (optional but recommended to use a virtual environment):

   • python -m venv venv
   • source venv/bin/activate (Linux/macOS) or venv\Scripts\activate (Windows)

2. Install the MCP server package in development mode (as per the project README):

   • pip install -e .

3. Configure environment variables before running:

   • ELASTIC_HOST=http://localhost:9200
   • ELASTIC_USERNAME=your_username
   • ELASTIC_PASSWORD=your_password
   • MCP_PORT=9999 (optional; defaults to 9999 if not set)

4. Run the server (by module name defined in your package):

   • python -m elasticsearch7_mcp_server

If you prefer a quick start, you can also explore running via Smithery if the project supports it, as described in the README: use npx -y @smithery/cli install @imlewc/elasticsearch7-mcp-server --client claude to install for Claude Desktop, though this is optional for local development.

Tips and common issues:

• Ensure the Elasticsearch cluster is reachable at ELASTIC_HOST and that the credentials (ELASTIC_USERNAME, ELASTIC_PASSWORD) are correct.
• If MCP_PORT is in use, choose an available port or stop the conflicting service.
• When using Docker Compose, ensure the environment file (.env) defines ELASTIC_PASSWORD and that ELASTIC_HOST points to the correct Elasticsearch endpoint.
• The es-search endpoint supports a variety of query structures, including full-text search, term queries, range filters, and aggregations. Refer to the README examples for building complex queries.
• If you encounter connection issues, verify network access between the MCP server and Elasticsearch and check Elasticsearch cluster health.
• The server runs on Python; avoid mixing incompatible Python environments if using virtual environments.