Elasticsearch MCP Server

@sunilemanjee

代码仓库

基本信息

分类

数据库

传输方式

stdio

作者

sunilemanjee

关于 Elasticsearch MCP Server

暂无概览

Elasticsearch MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "Elastic-Python-MCP-Server": {
      "command": "python3",
      "args": [
        "-m",
        "venv",
        "venv"
      ]
    }
  }
}

Elasticsearch MCP Server 工具

工具列表暂未提供

当前收录暂时没有可展示的工具列表，不代表服务没有工具。请查阅项目文档了解支持的能力。

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is Elasticsearch MCP Server?

Elasticsearch MCP Server is a Python‑based MCP (Model Context Protocol) server that provides an interface for searching and analyzing property data using Elasticsearch. It was converted from a Jupyter notebook and connects to Elasticsearch Serverless, using ELSER for semantic search and Google Maps API for geocoding.

How to use Elasticsearch MCP Server?

Clone the repository and copy env_config.template.sh to env_config.sh.
Set environment variables (ES_URL, ES_API_KEY, GOOGLE_MAPS_API_KEY, etc.) in env_config.sh.
Create a Python virtual environment, install dependencies from requirements.txt, and source the config file.
Run ./run_server.sh; the server starts on port 8001 by default.

Key features of Elasticsearch MCP Server

Property search by location, price, bedrooms, bathrooms, square footage, features, tax, and maintenance.
Geocoding integration using Google Maps API.
Connects to Elasticsearch Serverless with ELSER semantic search.
Supports custom search templates.
Provides three MCP tools: get_properties_template_params, geocode_location, search_template.

Use cases of Elasticsearch MCP Server

Searching a property database with natural language queries.
Converting location strings into geographic coordinates for proximity searches.
Enabling LLMs to retrieve structured property data via MCP.
Real‑estate analysis and portfolio management.

常见问题

What are the prerequisites?

Python 3.x, an Elasticsearch Serverless instance, a Google Maps API key (with Geocoding and Maps JavaScript APIs enabled), and the required Python packages.

How do I configure environment variables?

Copy `env_config.template.sh` to `env_config.sh` and set `ES_URL`, `ES_API_KEY`, `GOOGLE_MAPS_API_KEY`, and optionally `PROPERTIES_SEARCH_TEMPLATE`, `ELSER_INFERENCE_ID`, `ES_INDEX`, and `MCP_PORT`.

How do I run the server?

Source the environment variables (`source env_config.sh`) and execute `./run_server.sh`. Verify with `curl -v http://localhost:8001/sse`.

Can I restrict the Elasticsearch API key to read‑only?

Yes. After ingestion, create or edit an API key with privileges limited to `monitor` cluster and `read`/`view_index_metadata` on `properties` and `properties_raw` indices.

What MCP endpoints are available?

Three tools: `get_properties_template_params` (returns search template parameters), `geocode_location` (converts location to coordinates), and `search_template` (performs property searches).