MCP Elasticsearch Server 🔍

What is MCP Elasticsearch Server 🔍?

MCP Elasticsearch Server 🔍 is an MCP server that provides Elasticsearch integration capabilities to AI assistants and other MCP clients. It enables querying Elasticsearch clusters through a standardized interface.

How to use MCP Elasticsearch Server 🔍?

Install by cloning the repository, building with Go, and installing the binary. Configure with the ES_URL environment variable and either an ES_API_KEY or ES_USERNAME/ES_PASSWORD. Run directly or integrate with clients like Claude Desktop or ApMentor by adding a mcpServers entry.

Key features of MCP Elasticsearch Server 🔍

• Multiple authentication methods (API key or username/password)
• List indices with health status and document counts
• Retrieve field mappings to understand index structure
• Execute complex Elasticsearch queries with aggregations and sorting
• JSON-formatted output with search metadata
• Query execution time tracking

Use cases of MCP Elasticsearch Server 🔍

• List all or filtered Elasticsearch indices
• Get field mappings for one or more indices
• Execute simple or complex search queries across indices
• Search with aggregations and sorting for analytics
• Filter log indices by pattern for targeted searching

FAQ from MCP Elasticsearch Server 🔍

What authentication methods does MCP Elasticsearch Server 🔍 support?

It supports both API key authentication and basic authentication (username/password). You must provide either an API key or both username and password.

What are the runtime requirements?

The server is a Go binary that requires a running Elasticsearch cluster accessible via the provided URL.

Where does my data live?

All data resides in your Elasticsearch cluster. The server only queries and returns results; it does not store any data.

Are there any limitations on search results?

The default maximum number of documents returned is 10, with a configurable maximum of 10,000. You can also specify sorting and track total hits.

How does the server handle errors?

Detailed error messages are provided for authentication failures, index not found, query syntax errors, and connection issues.