ModelsAtlas
Explore
Atlas / MCP servers / rileylemm / GraphRAG

GraphRAGSAFE

mcp/rileylemm/graphrag
MCP server by rileylemm · Audited 2026-10-07
⚑ Report / re-scanSource ↗

This is a MCP server I built to interact with my hybrid graph rag db.

Verdict
SAFE
Grade
B
Trust score
89 /100
Exposed tools
2 2r · 0w · 0d
Transport
—
License
MIT
Stars
64
01

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

A Model Context Protocol server for querying a hybrid graph and vector database system, combining Neo4j (graph database) and Qdrant (vector database) for powerful semantic and graph-based document retrieval.

Overview

GraphRAG MCP provides a seamless integration between large language models and a hybrid retrieval system that leverages the strengths of both graph databases (Neo4j) and vector databases (Qdrant). This enables:

  • Semantic search through document embeddings
  • Graph-based context expansion following relationships
  • Hybrid search combining vector similarity with graph relationships
  • Full integration with Claude and other LLMs through MCP

This project follows the Model Context Protocol specification, making it compatible with any MCP-enabled client.

Features

  • Semantic search using sentence embeddings and Qdrant
  • Graph-based context expansion using Neo4j
  • Hybrid search combining both approaches
  • MCP tools and resources for LLM integration
  • Full documentation of Neo4j schema and Qdrant collection information

Prerequisites

  • Python 3.12+
  • Neo4j running on localhost:7687 (default configuration)
  • Qdrant running on localhost:6333 (default configuration)
  • Document data indexed in both databases

Installation

Quick Start

  1. Clone this repository:
git clone https://github.com/rileylemm/graphrag_mcp.git
cd graphrag_mcp
  1. Install dependencies with uv:
uv install
  1. Configure your database connections in the .env file:
# Neo4j Configuration
NEO4J_URI=bolt://localhost:7687
NEO4J_USER=neo4j
NEO4J_PASSWORD=password

# Qdrant Configuration
QDRANT_HOST=localhost
QDRANT_PORT=6333
QDRANT_COLLECTION=document_chunks
  1. Run the server:
uv run main.py

Detailed Setup Guide

For a detailed guide on setting up the underlying hybrid database sys

Read from source at commit 1a1092658091OBSERVED · 2026-10-07
02

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code
claude mcp add graphrag-mcp --env NEO4J_PASSWORD=${NEO4J_PASSWORD} -- uvx graphrag-mcp
claude-desktop
{
  "mcpServers": {
    "graphrag-mcp": {
      "command": "uvx",
      "args": [
        "graphrag-mcp"
      ],
      "env": {
        "NEO4J_PASSWORD": "${NEO4J_PASSWORD}"
      }
    }
  }
}
03

Exposed tools (2)

2 read · 0 write · 0 destructive.

ToolRiskDescription
hybrid_searchread
search_documentationread
04

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codePASS
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfacePASS
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
none-observed
Network
none-observed
Shell
none-observed
Dependencies
pinned
Secrets in source
none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.

Audited 2026-10-07 · audit v0.4.1 · source sha 1a1092658091full audit observations/trust-audit/mcp-server/rileylemm__graphrag.json · Report an issue / request a re-scan
05

Audit history

Every audit this server has had. A grade with a past is a grade somebody is still checking.

DateSourceVerdictGradeScoreChange
2026-10-071a1092658091SAFEB89first audit
06

Questions

What is the GraphRAG MCP server?

This is a MCP server I built to interact with my hybrid graph rag db.

What tools does GraphRAG expose?

2 in total: 2 read-only, 0 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.

Is GraphRAG safe to connect to an agent?

The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.

What credentials does GraphRAG need?

It reads NEO4J_PASSWORD from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.

How current is this page?

The grade is for one exact copy of the source (1a1092658091), read on 2026-10-07. The repository is watched and re-audited when it changes.

Advertisement