ArangoDB MCP Server for Python

A production-ready Model Context Protocol (MCP) server exposing advanced ArangoDB operations to AI assistants like Claude Desktop and Augment Code. Features async-first Python architecture, comprehensive graph management, flexible content conversion (JSON, Markdown, YAML, Table), backup/restore functionality, and analytics capabilities.

Quick Links

📚 Documentation: https://github.com/PCfVW/mcp-arango-async/tree/master/docs

🚀 Quick Start: https://github.com/PCfVW/mcp-arango-async/blob/master/docs/getting-started/quickstart.md

🔧 ArangoDB Setup: https://github.com/PCfVW/mcp-arango-async/blob/master/docs/getting-started/install-arangodb.md

🗄️ Multi-Tenancy Guide: https://github.com/PCfVW/mcp-arango-async/blob/master/docs/user-guide/multi-tenancy-guide.md

⚙️ CLI Reference: https://github.com/PCfVW/mcp-arango-async/blob/master/docs/user-guide/cli-reference.md

📖 Tools Reference: https://github.com/PCfVW/mcp-arango-async/blob/master/docs/user-guide/tools-reference.md

🎯 MCP Design Patterns: https://github.com/PCfVW/mcp-arango-async/blob/master/docs/user-guide/mcp-design-patterns.md

📝 Changelog: https://github.com/PCfVW/mcp-arango-async/blob/master/docs/developer-guide/changelog.md

🐛 Issues: https://github.com/PCfVW/mcp-arango-async/issues

Features

• ✅ 46 MCP Tools - Complete ArangoDB operations (queries, collections, indexes, graphs)
• ✅ Multi-Tenancy - Work with multiple databases, environment switching, cross-database operations
• ✅ MCP Design Patterns - Progressive discovery, context switching, tool unloading (98.7% token savings)
• ✅ Graph Management - Create, traverse, backup/restore named graphs
• ✅ Content Conversion - JSON, Markdown, YAML, and Table formats
• ✅ Backup/Restore - Collection and graph-level backup with validation
• ✅ Analytics - Query profiling, explain plans, graph statistics
• ✅ Dual Transport - stdio (desktop clients) and HTTP (web/containerized)
• ✅ Docker Support - Run in Docker for isolation and reproducibility
• ✅ Production-Ready - Retry logic, graceful degradation, comprehensive error handling
• ✅ Type-Safe - Pydantic validation for all tool arguments

Architecture

┌────────────────────┐      ┌─────────────────────┐       ┌──────────────────┐
│   MCP Client       │      │  ArangoDB MCP       │       │   ArangoDB       │
│ (Claude, Augment)  │─────▶│  Server (Python)    │─────▶│  (Docker)        │
│                    │      │  • 46 Tools         │       │  • Multi-Model   │
│                    │      │  • Multi-Tenancy    │       │  • Graph Engine  │
│                    │      │  • Graph Mgmt       │       │  • AQL Engine    │
│                    │      │  • MCP Patterns     │       │                  │
└────────────────────┘      └─────────────────────┘       └──────────────────┘

Getting Started with ArangoDB

Prerequisites

• Docker and Docker Compose installed
• Python 3.11+ (for mcp-arangodb-async)

Step 1: Install ArangoDB

Create a docker-compose.yml file:

services:
  arangodb:
    image: arangodb:3.11
    environment:
      ARANGO_ROOT_PASSWORD: ${ARANGO_ROOT_PASSWORD:-changeme}
    ports:
      - "8529:8529"
    volumes:
      - arangodb_data:/var/lib/arangodb3
      - arangodb_apps:/var/lib/arangodb3-apps
    healthcheck:
      test: arangosh --server.username root --server.password "$ARANGO_ROOT_PASSWORD" --javascript.execute-string "require('@arangodb').db._version()" > /dev/null 2>&1 || exit 1
      interval: 5s
      timeout: 2s
      retries: 30
    restart: unless-stopped

volumes:
  arangodb_data:
    driver: local
  arangodb_apps:
    driver: local

Create a .env file:

# ArangoDB root password
ARANGO_ROOT_PASSWORD=changeme

# MCP Server connection settings
ARANGO_URL=http://localhost:8529
ARANGO_DB=mcp_arangodb_test
ARANGO_USERNAME=mcp_arangodb_user
ARANGO_PASSWORD=mcp_arangodb_password

Start ArangoDB:

docker compose --env-file .env up -d

Step 2: Install mcp-arangodb-async

Install the MCP server package:

pip install mcp-arangodb-async

# Create environment and install
conda create -n mcp-arango python=3.11
conda activate mcp-arango
pip install mcp-arangodb-async

# Or with mamba/micromamba:
# mamba create -n mcp-arango python=3.11
# mamba activate mcp-arango
# pip install mcp-arangodb-async

# Create environment and install
uv venv .venv --python 3.11
uv pip install mcp-arangodb-async

Step 3: Create Database and User

Create the database and user for the MCP server:

maa db add mcp_arangodb_test \
  --url http://localhost:8529 \
  --with-user mcp_arangodb_user \
  --env-file .env

Expected output:

The following actions will be performed:
  [ADD] Database 'mcp_arangodb_test'
  [ADD] User 'mcp_arangodb_user' (active: true)
  [GRANT] Permission rw: mcp_arangodb_user → mcp_arangodb_test

Are you sure you want to proceed? [y/N]: y
db add:
[ADDED] Database 'mcp_arangodb_test'
[ADDED] User 'mcp_arangodb_user' (active: true)
[GRANTED] Permission rw: mcp_arangodb_user → mcp_arangodb_test

Verify the database connection:

# Set environment variables
export ARANGO_URL=http://localhost:8529
export ARANGO_DB=mcp_arangodb_test
export ARANGO_USERNAME=mcp_arangodb_user
export ARANGO_PASSWORD=mcp_arangodb_password

# Run health check
maa health

Expected output:

{"status": "healthy", "database_connected": true, "database_info": {"version": "3.11.x", "name": "mcp_arangodb_test"}}

Step 4: Configure MCP Host

Configure your MCP host to use the server. The configuration includes environment variables for database connection. The location of the configuration file depends on your MCP host. For Claude Desktop, the file is located at:

• Windows: %APPDATA%\Claude\claude_desktop_config.json
• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

Configuration:

{
  "mcpServers": {
    "arangodb": {
      "command": "python",
      "args": ["-m", "mcp_arangodb_async"],
      "env": {
        "ARANGO_URL": "http://localhost:8529",
        "ARANGO_DB": "mcp_arangodb_test",
        "ARANGO_USERNAME": "mcp_arangodb_user",
        "ARANGO_PASSWORD": "mcp_arangodb_password"
      }
    }
  }
}

{
  "mcpServers": {
    "arangodb": {
      "command": "conda",
      "args": ["run", "-n", "mcp-arango", "maa", "server"],
      "env": {
        "ARANGO_URL": "http://localhost:8529",
        "ARANGO_DB": "mcp_arangodb_test",
        "ARANGO_USERNAME": "mcp_arangodb_user",
        "ARANGO_PASSWORD": "mcp_arangodb_password"
      }
    }
  }
}

{
  "mcpServers": {
    "arangodb": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/project", "maa", "server"],
      "env": {
        "ARANGO_URL": "http://localhost:8529",
        "ARANGO_DB": "mcp_arangodb_test",
        "ARANGO_USERNAME": "mcp_arangodb_user",
        "ARANGO_PASSWORD": "mcp_arangodb_password"
      }
    }
  }
}

Restart your MCP client after updating the configuration.

Test the connection:

Ask your MCP client: "List all collections in the ArangoDB database"

The assistant should successfully connect and list your collections.

Available Tools

The server exposes 46 MCP tools organized into 11 categories:

Multi-Tenancy Tools (4 tools)

• arango_set_focused_database - Set focused database for session
• arango_get_focused_database - Get currently focused database
• arango_list_available_databases - List all configured databases
• arango_get_database_resolution - Show database resolution algorithm