LABSCO SUMMARY

The analyzer question gets answered in the same session as the query that raised it.

"Why doesn't this document match?" is usually a tokenization question, and analyze_text answers it directly instead of by inference from result sets, while general_api_request keeps anything the dedicated tools skip one HTTP shape away. Deletes sit in the same tool set as the reads and take no confirmation step, and dropping a data stream takes its backing indices with it — so before pointing this at production, either set DISABLE_HIGH_RISK_OPERATIONS=true or name the calls you want gone in DISABLE_OPERATIONS.

What it is

A query and administration surface over Elasticsearch and OpenSearch. Every tool takes an optional cluster name, so several named clusters can be addressed in one session.

What you get

Document work end to end: search with a query body, create or update with an optional id, fetch or delete by id, and delete everything matching a query
Index lifecycle — list them, read mappings, settings and aliases for one, create with a configuration body, delete
Aliases as their own tools rather than a raw call: list all, read one index's aliases, create or update, delete
Data streams created, inspected by name or wildcard, and deleted along with their backing indices
Cluster health and high-level statistics
analyze_text runs a string through an analyzer, tokenizer, filter or char_filter — or through an index's configured analyzer — and shows the tokens that come out
general_api_request takes a method, path, params and body, so an endpoint with no dedicated tool is still reachable
Write tools removed from the client's view entirely with DISABLE_HIGH_RISK_OPERATIONS=true, or a named subset removed with DISABLE_OPERATIONS
RESPONSE_FORMAT=gcf re-encodes tool results as Graph Compact Format — measured at about 39% fewer tokens than compact JSON on representative responses — while structured output is left as JSON

Requirements

A reachable cluster and uv, since it starts through uvx. Hosts and credentials come from the environment: ELASTICSEARCH_HOSTS with ELASTICSEARCH_USERNAME and ELASTICSEARCH_PASSWORD, or ELASTICSEARCH_API_KEY, and the OPENSEARCH_ equivalents for OpenSearch; several named clusters go in ELASTICSEARCH_CLUSTERS or a file named by ELASTICSEARCH_CLUSTERS_FILE, with DEFAULT_CLUSTER deciding which one a call without a cluster lands on. Pick the package that matches the server: elasticsearch-mcp-server for 8.x, elasticsearch-mcp-server-es7 for 7.x, elasticsearch-mcp-server-es9 for 9.x, opensearch-mcp-server for OpenSearch. Over SSE or Streamable HTTP, set MCP_API_KEY — without it the server answers anyone who can reach the port. Apache-2.0.

Setup effort

One command plus credentials — uvx elasticsearch-mcp-server, then supply credentials

See installation options

USE CASES

Working out why a document that obviously matches does not come back

Run the text through the index's analyzer and look at the tokens it actually produces, then compare that against the mapping and the query you were running.

analyze_textget_indexsearch_documents

Standing up an index for something new

Create the index with its mappings and settings in the body, put an alias in front of it so readers never name the concrete index, and write the first documents.

create_indexput_aliasindex_documentget_document

Cutting readers over to a rebuilt index

Read what the alias points at now, repoint it at the new index, and drop the alias that is no longer wanted — nothing downstream has to learn a new name.

list_aliasesget_aliasput_aliasdelete_alias

Checking on a cluster that has changed colour

Health first, then the statistics overview, then the index list — enough to tell an unassigned-shard problem from a disk problem before opening a dashboard.

get_cluster_healthget_cluster_statslist_indices

Comparing staging against production without reconnecting

Every call takes a cluster name and falls back to the default one, so the same search or index inspection can be run against both in a single session.

list_indicessearch_documentsget_index

Reaching an API the dedicated tools do not cover

Send the method, path, query params and body yourself — reindex, ILM, whatever the twenty tools skip.

general_api_request

20 tools

Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.

`analyze_text`

Analyze text to see how it would be tokenized.

Use this tool to understand how Elasticsearch/OpenSearch tokenizes and transforms text using analyzers. This is essential for debugging search queries and understanding why certain documents match or don't match.

Args: text: The text to analyze index: Index name to use its configured analyzer. If not specified, uses cluster-level analysis with built-in analyzers only. analyzer: Name of the analyzer to use (e.g., 'standard', 'korean', 'korean_search'). If index is specified, you can use custom analyzers defined in that index. tokenizer: Tokenizer to use for custom analysis chain. Cannot be used together with 'analyzer'. filter: List of token filters to apply (e.g., ['lowercase', 'stop']). Used with 'tokenizer' for custom analysis chain. char_filter: List of character filters to apply before tokenization. Used with 'tokenizer' for custom analysis chain. explain: If True, returns detailed information about each token including all token attributes and filter transformations. Useful for debugging complex analyzer chains. attributes: List of token attributes to return when explain=True (e.g., ['keyword', 'type']). If not specified, all attributes are returned. cluster: Optional cluster name. Uses the default cluster if omitted.

Returns: Dict containing 'tokens' array. Each token has 'token', 'start_offset', 'end_offset', 'type', and 'position' fields. With explain=True, returns detailed 'detail' object showing each filter's effect.

INSTALLATION

Connect your client

Server setup

TERMINAL

uvx elasticsearch-mcp-server

For Elasticsearch 8.x. On 7.x run elasticsearch-mcp-server-es7 instead, on 9.x elasticsearch-mcp-server-es9, and on OpenSearch opensearch-mcp-server. Set ELASTICSEARCH_HOSTS and your credentials in the environment first.

MCP Clients

Claude Desktop — password

CLAUDE DESKTOP — PASSWORD · JSON

{
  "mcpServers": {
    "elasticsearch-mcp-server": {
      "command": "uvx",
      "args": [
        "elasticsearch-mcp-server"
      ],
      "env": {
        "ELASTICSEARCH_HOSTS": "https://localhost:9200",
        "ELASTICSEARCH_USERNAME": "elastic",
        "ELASTICSEARCH_PASSWORD": "test123"
      }
    }
  }
}

Claude Desktop — API key

CLAUDE DESKTOP — API KEY · JSON

{
  "mcpServers": {
    "elasticsearch-mcp-server": {
      "command": "uvx",
      "args": [
        "elasticsearch-mcp-server"
      ],
      "env": {
        "ELASTICSEARCH_HOSTS": "https://localhost:9200",
        "ELASTICSEARCH_API_KEY": "<YOUR_ELASTICSEARCH_API_KEY>"
      }
    }
  }
}

Claude Desktop — OpenSearch

CLAUDE DESKTOP — OPENSEARCH · JSON

{
  "mcpServers": {
    "opensearch-mcp-server": {
      "command": "uvx",
      "args": [
        "opensearch-mcp-server"
      ],
      "env": {
        "OPENSEARCH_HOSTS": "https://localhost:9200",
        "OPENSEARCH_USERNAME": "admin",
        "OPENSEARCH_PASSWORD": "admin"
      }
    }
  }
}

Any MCP client — clusters inline

ANY MCP CLIENT — CLUSTERS INLINE · JSON

{
  "mcpServers": {
    "elasticsearch-mcp-server": {
      "command": "uvx",
      "args": [
        "elasticsearch-mcp-server"
      ],
      "env": {
        "ELASTICSEARCH_CLUSTERS": "{\"prod\": {\"hosts\": [\"https://prod-es:9200\"], \"api_key\": \"<YOUR_PROD_API_KEY>\", \"verify_certs\": true}, \"staging\": {\"hosts\": [\"https://staging-es:9200\"], \"username\": \"elastic\", \"password\": \"<YOUR_STAGING_PASSWORD>\"}}",
        "DEFAULT_CLUSTER": "prod"
      }
    }
  }
}

Any MCP client — clusters file

ANY MCP CLIENT — CLUSTERS FILE · JSON

{
  "mcpServers": {
    "elasticsearch-mcp-server": {
      "command": "uvx",
      "args": [
        "elasticsearch-mcp-server"
      ],
      "env": {
        "ELASTICSEARCH_CLUSTERS_FILE": "/etc/mcp/es-clusters.json",
        "DEFAULT_CLUSTER": "prod"
      }
    }
  }
}

SSE server

Export the cluster connection and a key that locks the server, then start it. It listens on http://127.0.0.1:8000/sse by default; add --host, --port and --path only if you need another address, and keep MCP_API_KEY set whenever it is reachable from other machines:

SSE SERVER · TERMINAL

export ELASTICSEARCH_HOSTS="https://localhost:9200"
export ELASTICSEARCH_API_KEY="<YOUR_ELASTICSEARCH_API_KEY>"
export MCP_API_KEY=$(openssl rand -base64 32)
uvx elasticsearch-mcp-server --transport sse

Point your client at http://127.0.0.1:8000/sse and send the header Authorization: Bearer <YOUR_MCP_API_KEY>.

HTTP server

Export the cluster connection and a key that locks the server, then start it. It listens on http://127.0.0.1:8000/mcp by default; add --host, --port and --path only if you need another address, and keep MCP_API_KEY set whenever it is reachable from other machines:

HTTP SERVER · TERMINAL

export ELASTICSEARCH_HOSTS="https://localhost:9200"
export ELASTICSEARCH_API_KEY="<YOUR_ELASTICSEARCH_API_KEY>"
export MCP_API_KEY=$(openssl rand -base64 32)
uvx elasticsearch-mcp-server --transport streamable-http

Point your client at http://127.0.0.1:8000/mcp and send the header Authorization: Bearer <YOUR_MCP_API_KEY>.

Elasticsearch & OpenSearch MCP

The analyzer question gets answered in the same session as the query that raised it.

What it is

What you get

Requirements

Setup effort

Working out why a document that obviously matches does not come back

Standing up an index for something new

Cutting readers over to a rebuilt index

Checking on a cluster that has changed colour

Comparing staging against production without reconnecting

Reaching an API the dedicated tools do not cover

20 tools

analyze_text

create_data_stream

create_index

delete_alias

delete_by_query

delete_data_stream

delete_document

delete_index

general_api_request

get_alias

get_cluster_health

get_cluster_stats

get_data_stream

get_document

get_index

index_document

list_aliases

list_indices

put_alias

search_documents

What to know first

Tools you get

Requires

Light Standing cost

Connect your client

Server setup

MCP Clients

Claude Desktop — password

Claude Desktop — API key

Claude Desktop — OpenSearch

Any MCP client — clusters inline

Any MCP client — clusters file

SSE server

HTTP server

Same category

Couchbase

Elastic MCP

ArangoDB

InfluxDB MCP Server

`analyze_text`

`create_data_stream`

`create_index`

`delete_alias`

`delete_by_query`

`delete_data_stream`

`delete_document`

`delete_index`

`general_api_request`

`get_alias`

`get_cluster_health`

`get_cluster_stats`

`get_data_stream`

`get_document`

`get_index`

`index_document`

`list_aliases`

`list_indices`

`put_alias`

`search_documents`