AR
Data Analytics

ArangoDB Graph MCP Server — Async Python Analytics

Deploy an ArangoDB Graph MCP server with async Python, graph management, flexible JSON/Markdown/YAML conversion, backup/restore and analytics for AI apps

View on GitHub
Quick Install
npx -y @PCfVW/mcp-arangodb-async

Overview

ArangoDB Graph MCP Server — Async Python Analytics is an async-first Model Context Protocol (MCP) server that exposes advanced ArangoDB operations to AI assistants and automation tools. It packages 46 MCP tools for collection, index, AQL and graph management, plus backup/restore, content conversion (JSON, Markdown, YAML, table), profiling and analytics. The server is built for production usage: multi-tenancy, retry logic, type-safe tool arguments (Pydantic), and dual transports (stdio for desktop clients and HTTP for containerized/web use).

This server is useful when you need programmatic graph analytics and robust data management integrated into AI workflows: feeding semantically-organized context into LLMs, running explain/profile cycles on AQL, maintaining reproducible backups for graph datasets, or exposing curated graph operations to assistants such as Claude or Augment.

Repository: https://github.com/PCfVW/mcp-arangodb-async Tags: data-analytics

Features

  • Async-first Python architecture (Python 3.11+)
  • 46 MCP tools covering:
    • Collections, indexes, documents
    • AQL queries, explain and profile
    • Named graph create/traverse/delete
    • Backup/restore (collection and graph)
    • Content conversion: JSON ↔ Markdown ↔ YAML ↔ Table
  • Multi-tenancy and environment switching
  • Dual transport: stdio (desktop MCP clients) and HTTP (web/container)
  • Docker support for ArangoDB and reproducible deployments
  • Pydantic validation for all tool inputs
  • Production-minded: retry logic, graceful degradation, error handling
  • Analytics: query profiling, execution plans, graph statistics

Installation / Configuration

Prerequisites:

  • Docker & Docker Compose
  • Python 3.11+
  1. Start ArangoDB (docker-compose)

Create docker-compose.yml:

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
    restart: unless-stopped

volumes:
  arangodb_data:
  arangodb_apps:

Create .env:

ARANGO_ROOT_PASSWORD=changeme
ARANGO_URL=http://localhost:8529
ARANGO_DB=mcp_arangodb_test
ARANGO_USERNAME=mcp_arangodb_user
ARANGO_PASSWORD=mcp_arangodb_password

Start the container:

docker compose --env-file .env up -d
  1. Install the MCP server package
pip install mcp-arangodb-async
  1. Configure and run the MCP server

The package exposes a server component (see repository docs/CLI reference for exact runtime flags). Typical configuration uses environment variables shown above to point the server to ArangoDB. The server supports both HTTP and stdio transports; choose the mode that fits your client integration (webhooks vs local assistant bridge).

Example (illustrative):

# run as a module (example)
python -m mcp_arangodb_async --transport http --host 0.0.0.0 --port 8080

For production deployments, run the process in a container or process manager and bind only necessary ports. Consult the GitHub docs for full CLI and container examples: https://github.com/PCfVW/mcp-arangodb-async/tree/master/docs

Available Tools / Resources

The server provides 46 tools across these categories. Example tool names and capabilities:

  • Collections & Documents
    • create-collection, drop-collection, list-collections
    • insert-document, update-document, delete-document
  • Indexes
    • create-index, list-indexes, drop-index
  • AQL & Analytics
    • query-execute, query-explain, query-profile, query-kill
  • Graph Management
    • create-graph, drop-graph, traverse-graph, graph-stats
  • Backup & Restore
    • backup-collection, restore-collection, backup-graph, restore-graph
  • Conversion
    • convert-to-json, convert-to-markdown, convert-to-yaml, convert-to-table

All tool inputs are validated with Pydantic models and return structured results suitable for downstream LLM consumption or orchestration pipelines.

Use Cases (concrete examples)

  • AI-assisted context retrieval
    • Use graph traversal and convert-to-markdown to feed concise, context-rich passages to an LLM assistant that composes answers or code.
  • Query diagnostics
    • Execute query-explain and query-profile before running heavy AQL jobs; feed explain output into an optimizer assistant for suggested index changes.
  • Backup & restore workflows
    • Schedule periodic backup-graph and validate backups. Restore into a temporary workspace for safe experimentation by an AI agent.
  • Multi-tenant analytics
    • Single MCP server manages multiple databases for isolated teams; tools can switch environments and perform cross-database analytics.
  • ETL & content conversion
    • Import documents, transform them to YAML/Markdown for knowledge-base ingestion, and export tabular snapshots for reporting.
  • Quickstart & detailed usage examples: https://github.com/PCfVW/mcp-arangodb-async/tree/master/docs/getting-started
  • CLI and tools reference: https://github.com/PCfVW/mcp-arangodb-async/tree/master/docs/user-guide
  • Design patterns for MCP integrations: docs/user-guide/mcp-design-patterns.md in the repo
  • Issues & contribution: https://github.com/PCfVW/mcp-arangodb-async/issues

This server is a pragmatic bridge between ArangoDB’s multi-model graph capabilities and AI-driven workflows — designed

Tags:data-analytics