# Gridscale MCP > Gridscale Cloud Hosting MCP. Manage your entire cloud infrastructure—servers, networking, databases, storage volumes—directly through natural language prompts in any AI agent. Provision new compute nodes, check real-time CPU usage, and manage firewall rules without leaving your chat interface. ## Overview - **Category:** developer-tools - **Price:** Free - **Tags:** iaas, paas, cloud-hosting, server-management, infrastructure-as-code ## Description Need to make a change on your cloud setup? You don't have to log into five different dashboards or remember complex CLI syntax just to audit a few resources. This MCP connects your AI agent directly to Gridscale’s entire suite of IaaS and PaaS tools. You talk about what you need—like starting up a staging server or checking the CPU load on production nodes—and it handles the API calls, giving you immediate status reports. It lets you manage everything from creating new virtual machines and linking storage volumes to listing available IP addresses and managing firewall rules, all through conversation. If your team uses Vinkius for its centralized catalog of MCPs, this connector makes sure cloud infrastructure management is just another task you delegate to your agent. ## Tools ### create_server Builds and provisions an entirely new cloud server instance for your account. ### get_server_metrics Pulls performance data, like CPU usage or IOPS, for a specific running server. ### get_server_power Checks the current power status of any given server (on, off, etc.). ### link_ip_to_server Attaches a specific IP address to a running or stopped cloud server. ### link_storage_to_server Connects an existing storage volume to a server, which usually requires the server to be powered off first. ### list_deleted Retrieves historical records of objects that have been marked for deletion. ### list_firewalls Shows a list of all existing firewall rules and configurations. ### list_ips Gives you an inventory of every IP address assigned to your account. ### list_iso_images Retrieves a list of all available operating system ISO images for deployment. ### list_locations Shows the datacenters or physical locations where you can provision new cloud resources. ### list_networks Lists all defined private networks within your Gridscale account. ### list_paas Provides an inventory of managed platform services like databases and Redis instances. ### list_servers Returns a full list of every server currently registered in your Gridscale account. ### list_storages Lists all available and existing storage volumes you own. ### list_templates Shows a list of pre-defined machine images or templates for quick resource deployment. ### set_server_power Changes the power status of a server, setting it to on or off. ### shutdown_server Initiates a proper ACPI shutdown procedure for a specified cloud server. ## Prompt Examples **Prompt:** ``` List all my active servers in Gridscale. ``` **Response:** ``` I've retrieved your server list. You have 3 servers: 'Production-Web' (Online), 'Staging-App' (Online), and 'Legacy-DB' (Offline). Would you like to see the metrics for any of them? ``` **Prompt:** ``` Create a new server named 'Dev-Node' with 2 cores and 4GB RAM. ``` **Response:** ``` I am initiating the creation of 'Dev-Node' with 2 CPU cores and 4GB RAM. I'll need a location UUID to proceed—should I list the available locations for you? ``` **Prompt:** ``` Show me the performance metrics for server uuid-789-xyz. ``` **Response:** ``` Fetching metrics for server uuid-789-xyz... In the last 15 minutes, CPU usage was stable at 15% and storage IOPS averaged 320. The server is performing within normal parameters. ``` ## Capabilities ### Provisioning & Lifecycle Management Create new compute servers or change their power state (on/off) through simple commands. ### Performance Monitoring Get real-time metrics, including CPU usage and storage IOPS, for specific servers to assess health. ### Network Control Manage IP addresses, private networks, and firewall rules associated with your cloud resources. ### Resource Linking & Storage Connect storage volumes or ISO images to servers when they're ready for deployment. ### Platform Service Access (PaaS) List and manage managed services like databases, Redis instances, and other platform components. ## Use Cases ### The Weekend Audit A system administrator needs to check if all 14 staging servers are correctly configured before the Monday release. They prompt their agent: 'List all servers' and then follow up with, 'For those listed, get_server_metrics for CPU usage.' The agent compiles the list and provides a single summary of resource health. ### The New Dev Environment A developer needs a clean test machine. They ask the agent to create_server using a specific template, then immediately run list_networks to ensure it lands in the correct private network segment before linking resources. ### Fixing a Broken Service A production service is unreachable after an update. The engineer asks the agent to check network connectivity by listing all firewalls, identifying which rule broke the port, and then instructing the agent to modify or remove the incorrect rule. ### Decommissioning Hardware The team is retiring old hardware. Instead of manually navigating multiple screens, they prompt the agent to list_servers for all nodes marked 'legacy' and then issue a shutdown_server command on each one sequentially. ## Benefits - Instant Auditing: Need to know which servers are running or how much CPU a specific node is using? Just ask the agent. It pulls metrics instantly, saving you from jumping between monitoring dashboards and your chat window. - Safe Deployment: Don't forget to check prerequisites before deployment. Use list_storages and list_iso_images first. Then, use link_storage_to_server only after confirming the server is powered off via get_server_power. - Network Visibility: Managing IPs and firewalls used to be a nightmare of separate screens. Now, you can ask the agent to list_ips or list_firewalls and understand your network perimeter in one chat exchange. - Lifecycle Control: Forget manual shutdowns. You can issue an ACPI shutdown on a server using shutdown_server, then follow up by checking its status with get_server_power—all without leaving your workflow. - Rapid Provisioning: Need to test something? The agent handles the full flow: list_locations for the right region, then create_server, and finally link_ips to make it accessible. ## How It Works The bottom line is you tell your agent what infrastructure state you need, and it executes the required API calls across Gridscale’s services. 1. Subscribe to this MCP using your Gridscale User ID and API Token. 2. Connect the credentials to any compatible AI agent (like Claude or Cursor). 3. Use natural language prompts, like 'List all servers that are currently offline,' and receive actionable data back. ## Frequently Asked Questions **How do I check a server's performance using get_server_metrics?** You provide the specific server ID to the agent, and it pulls metrics like CPU usage or IOPS. The agent then reports if the machine is operating normally based on that data. **Do I need to shut down a server before linking storage with link_storage_to_server?** Yes, you do. You should first use get_server_power or shutdown_server to ensure the target server is powered off. The API call for linking storage often requires this prerequisite. **What is the best way to find a new location to deploy my servers?** Use list_locations first. This gives you all available datacenters or regions where you can reliably provision resources before calling create_server. **Can I check if a server is running using get_server_power?** Yes, that's the primary function. It gives you the current operational state—on or off—so you know exactly what power level to expect from any given ID. **When I use `create_server`, what key details do I need to provide besides just the CPU and RAM specifications?** You must specify a valid location UUID and often a template ID. The process requires these context points to build the server correctly, so always run `list_locations` first to get available options. **If I need to check existing security rules, how do I audit my current network settings using `list_firewalls`?** `list_firewalls` pulls up a complete list of all active firewalls and their associated rule sets. This lets you see exactly what traffic is currently allowed or blocked before making any changes. **I want to make sure I'm using the right IP address; how do I check if an IP is available for linking with `list_ips`?** `list_ips` retrieves every allocated IP in your account. Reviewing this list helps you verify availability and ensures that when you run `link_ip_to_server`, the target resource hasn't already been assigned. **How do I get a complete overview of all managed services, like databases or Redis instances, using `list_paas`?** `list_paas` gives you an inventory of every Platform as a Service resource attached to your account. This is the fastest way to see and manage database service statuses without guessing. **Can I check the power status of a specific server using its UUID?** Yes! Use the `get_server_power` tool with the server's UUID. The agent will return whether the server is currently powered on or off. **How do I monitor the performance of my cloud servers?** You can use the `get_server_metrics` tool. It retrieves data in 15-minute intervals for core usage and storage IOPS, allowing your AI to analyze resource consumption. **Does this integration support managed databases like MySQL or PostgreSQL?** Yes, you can list all managed Platform Services (PaaS) including MySQL, PostgreSQL, and Redis using the `list_paas` tool.