ElasticGraph MCP Server

This provides a Model Context Protocol (MCP) server for ElasticGraph using the MCP python-sdk.

Setup

1. Install dependencies:

# make install
uv sync

source .venv/bin/activate

2. Run the server:

The server runs on port 3000, and though there are no logs displayed, it is actively waiting for input.

# make server
uv pip install .
elasticgraph-mcp-server

Development

MCP Inspector

You can test your MCP server with Anthropic's Inspector. To start:

1. Run the following command, which starts the server as a subprocess and launches the Inspector UI:

# make inspector
mcp dev src/elasticgraph_mcp/server.py

2. Open your browser and navigate to http://localhost:5173 to access the MCP Inspector UI.

Goose

Add a development build to Goose:

1. In Goose, navigate to Settings > Extensions > Add.
2. Set Type to StandardIO.
3. Obtain the absolute path to the project’s CLI:

# Get the path and copy it to your clipboard
realpath .venv/bin/elasticgraph-mcp-server | pbcopy

4. Type uv run then paste that path

uv run </path/to/elasticgraph_mcp/.venv/bin/elasticgraph-mcp-server>

5. Enable the extension and verify that Goose recognizes your tools.

Ask goose: What tools and resources for ElasticGraph do you have?

Development commands

This project uses make for common development tasks. To see all available commands, run:

make help

Using Goose to help with development

You can use Goose to improve this MCP server. To teach goose about MCP, follow these steps:

Setting Up a Goose Session

1. Navigate to this MCP Server Directory:

cd ai_tools/elasticgraph-mcp-server

2. Prepare a Temporary Directory:

mkdir tmp
touch tmp/.gitignore
echo "/*" > tmp/.gitignore

3. Create MCP Instructions File:

Copy MCP instructions from MCP LLM instructions to a new file:

touch tmp/mcp_for_llm_instructions.md

4. Start a Goose Session:

goose session

Try this prompt:

First, learn about MCP servers from tmp/mcp_for_llm_instructions.md. Then, see the current MCP server I'm building in src/elasticgraph_mcp/server.py. Now