What features does Neo4j GDS MCP server provide?

Q: What features does Neo4j GDS MCP server provide?

GDS Algorithm Toolkit Integration: Includes toolings from Neo4j Graph Data Science library to run common graph algorithms.. Graph Question Answering: Allows asking questions about your Neo4j graph and receiving computed answers.. Parameterized Algorithm Execution via LLMs: LLMs equipped with GDS Agent can decide and execute the appropriate parameterized graph algorithms.. Multi-Algorithm Reasoning (Shortest Path + Yen's): Demonstrates using multiple algorithms in sequence to answer complex questions, e.g., travel planning.. London Underground Example Dataset: Provides an example dataset and workflow to load and query a graph in Neo4j using GDS.. Dev Start & Testing Workflow: Guides starting the MCP server in dev mode and running tests with pytest and code style checks.. Claude CLI Integration: Option to start Claude CLI with the GDS Agent for interactive usage.

Neo4j GDS MCP Server 2025: Features & Setup Guide

Overview

This MCP Server brings together the Neo4j Graph Data Science (GDS) toolkit with an MCP Agent, enabling LLM-powered graph reasoning over a live Neo4j database. It exposes the GDS library’s common graph algorithms so an AI agent can select and parameterize the appropriate method to answer graph questions. Once the server is running, you can ask questions about your graph and receive answers computed by the relevant algorithms, enabling tasks from path finding to centrality analysis. The GDS Agent demonstrates how an LLM equipped with this tool can orchestrate multiple algorithms (e.g., shortest path, Yen's K-Shortest Paths) to support complex planning and reasoning. The repository includes a a London Underground example dataset to illustrate loading, connecting to Neo4j, and querying the graph. Start commands for development are provided (uv sync --dev, uv run gds-agent or Claude CLI), and the server can be integrated with other MCP servers.

Features

GDS Algorithm Toolkit Integration

Includes toolings from Neo4j Graph Data Science library to run common graph algorithms.

Graph Question Answering

Allows asking questions about your Neo4j graph and receiving computed answers.

Parameterized Algorithm Execution via LLMs

LLMs equipped with GDS Agent can decide and execute the appropriate parameterized graph algorithms.

Multi-Algorithm Reasoning (Shortest Path + Yen's)

Demonstrates using multiple algorithms in sequence to answer complex questions, e.g., travel planning.

London Underground Example Dataset

Provides an example dataset and workflow to load and query a graph in Neo4j using GDS.

Dev Start & Testing Workflow

Guides starting the MCP server in dev mode and running tests with pytest and code style checks.

Claude CLI Integration

Option to start Claude CLI with the GDS Agent for interactive usage.