NebulaGraph

Open-source distributed graph database with a shared-nothing architecture, built for massive-scale graphs with millisecond latency.

MCPCLI1 Skill

About NebulaGraph

Open-source distributed graph database with a shared-nothing architecture, built for massive-scale graphs with millisecond latency. Explore how NebulaGraph integrates with the agentic data stack ecosystem and supports autonomous data operations.

Key Features

  • Shared-nothing distributed architecture with separate graphd, metad, and storaged services
  • Horizontal scalability to hundreds of billions of vertices and trillions of edges
  • nGQL query language with openCypher compatibility
  • Raft-based replication for strong consistency and high availability
  • Millisecond-latency multi-hop traversals for real-time workloads
  • Ecosystem tools: NebulaGraph Studio, Explorer, Exchange (Spark-based import), and Algorithm
  • Clients for Python, Java, Go, and C++
  • Community MCP server exposing schema, query, and shortcut graph algorithms to agents

Agent Integration

CLI — nebula-console

$ docker run --rm -ti vesoft/nebula-console (or download from GitHub releases)
CLI Documentation

Agent Skills

nebulagraph

External Links

NebulaGraph MCP Server

MCP server (pip install nebulagraph-mcp-server) for NebulaGraph 3.x schema discovery and queries

NebulaGraph Console

Native command-line client for running nGQL against NebulaGraph

NebulaGraph Agent Skill

Community skill for nGQL generation and schema design guidance via the MCP server

NebulaGraph GitHub

Main source repo for the distributed graph database

← Back to Graph Database