Integrates with Elastic products, providing semantic search capabilities via ELSER and kNN, hybrid search, and automatic index setup for Elasticsearch.
Supports managed Elastic Cloud clusters with API key authentication, enabling semantic search and index management.
Provides tools for interacting with Elasticsearch, including semantic search (ELSER, kNN), hybrid search, index management, and document indexing, optimized for Elasticsearch Serverless and Elastic Cloud.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Elasticsearch MCP Server Search for 'machine learning trends' using ELSER in my news index"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A FastAPI-based web server that exposes Elasticsearch semantic search and data retrieval capabilities through a simple REST API. This project connects directly to your Elasticsearch cluster (including Elastic Cloud and Serverless) and allows you to build custom search applications effortlessly.
Semantic Search: Use native Elasticsearch semantic capabilities directly through a REST API.
REST API First: Fully functional web API with built-in Swagger documentation.
Data Retrieval: Easily pull raw documents from any index.
Python Integration: Includes a helper script (multi_index_search.py) demonstrating how to interact with the API using Python.
Related MCP server: Meta MCP Search
Edit the .env file with your Elasticsearch connection details:
ELASTICSEARCH_HOSTS=https://your-cluster-url.es.aws.elastic.cloud:443
ELASTICSEARCH_API_KEY=your_api_key_hereStart the web server using the pre-configured executable in your virtual environment:
.\venv\Scripts\elasticsearch-api.exeOnce running, you can access:
API Health Check: http://localhost:8000/ping
Interactive Documentation (Swagger UI): http://localhost:8000/docs
A demo script (multi_index_search.py) is provided to show how to communicate with the API. While the server is running, open a new terminal and execute:
.\venv\Scripts\python.exe multi_index_search.pyThis script will:
Connect to the local API via http://localhost:8000/call
Retrieve records from the company_details index.
Perform native semantic searches on the om_poc_v2 index using natural language queries.
Print the matched products with their relevance scores, prices, categories, and descriptions!
GET /ping: Verify your cluster connection.
GET /tools: List all available internal tools.
POST /call: The primary dynamic endpoint. Pass a tool name (like es_raw_search) and its arguments (like index and query DSL body) to interact with the Elasticsearch cluster.
Python 3.11+
Elasticsearch 8.8+ (including Serverless)
Valid API Keys and a running cluster defined in your .env file.
This server cannot be deployed
Agent-driven search: build, import, tune, search, and score result quality — all over MCP.
Cloud or self-hosted knowledge for AI agents: hybrid search, reranking, GraphRAG, scoped MCP tools.
Make your knowledge agent-ready. One MCP endpoint, 5 connectors, 3 search modes.
