


















The IONOS CLOUD MCP Server is free to use. You pay only for the IONOS CLOUD resources your account already uses and there is no separate licence or subscription fee for the MCP Server itself. Build a tailored estimate for your specific requirements and production workloads with our dynamic pricing calculator.
Use case
Use your AI client to systematically inspect your IONOS CLOUD environment and surface misconfigurations before they become incidents.
Use case
Give your finance and engineering teams a natural-language interface to IONOS CLOUD billing data without writing a single script.
Use case
Connect AI agents to live IONOS CLOUD account data so they can reason over real infrastructure state rather than static documentation.
Yes. The IONOS CLOUD MCP Server is a free MCP server: there are no licence or subscription fees, and the source code is published under Apache 2.0 on GitHub. You are billed only for the IONOS CLOUD resources your account actually consumes. An IONOS CLOUD account and an API token are required.
Via Homebrew, as a pre-built binary for Linux, macOS and Windows, with go install, or as a Docker image. The container image is available on ghcr.io for linux/amd64 and linux/arm64, and runs locally, on a CI runner or inside a container. The server is also listed on the major MCP registries and IDE marketplaces, including the official MCP Registry, Smithery, mcp.so, the Cursor Directory and Glama. The full list is in the GitHub repository.
The MCP Server provides around 200 tools across seven products: Compute Engine, IONOS CLOUD Object Storage, Billing, Cloud DNS, Managed Kubernetes, Certificate Manager and Activity Log. That includes inspecting your Kubernetes clusters, node pools and available versions. Database as a Service with PostgreSQL, MongoDB and MariaDB is not covered at present. You can request further products via GitHub.
Not by default. Out of the box every tool is an inspection or query operation: write tools are not registered and do not appear in your AI client at all. Using them requires you to enable them explicitly. Until you do, connecting the server to live production accounts remains just as safe.
Yes. Create a separate token and give it only the permissions you actually need. For analysis and audit work, read access to Compute Engine, Cloud DNS, billing and Object Storage is enough. Write operations require broader permissions, so handle that token with more care and rotate it regularly.
The split matters: API calls go from your machine directly to IONOS CLOUD over HTTPS - no third-party relay involved. Your prompts and tool outputs are sent to whichever AI model provider your client uses (e.g. Anthropic, Microsoft). For a path where inference also stays within EU infrastructure, pair the MCP Server with the IONOS CLOUD AI Model Hub.
Dedicated setup guides cover Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Gemini CLI, JetBrains AI Assistant, Windsurf, Kiro, Continue, Cline, Zed, and OpenCode. Any other MCP-compliant client also works, including AI agent frameworks such as the OpenAI Agents SDK and LangGraph.
Yes. The full source code is published at github.com/ionos-cloud/ionoscloud-mcp under the Apache 2.0 licence. You can inspect exactly what the server does, contribute, or report issues directly via GitHub.