Cloud _server

The Cloud Services MCP server exposes a unified toolset for managing cloud resources across AWS, Azure, and GCP using the MCP protocol. It enables AI assistants to perform resource operations (list, create, delete, start, stop), analyze costs, monitor health and metrics, and run security checks through a consistent tool interface. You can connect your MCP clients (Cursor, Claude Desktop, or any MCP-compliant client) to the server by configuring an MCP server entry pointing to the Node.js server runtime, typically via node dist/index.js. Once connected, clients send MCP tool invocations to the server, which routes them to the appropriate cloud provider SDKs via adapters, returning structured results back to the client for decision-making.

You’ll have access to capabilities such as multi-cloud resource management (across AWS, Azure, and GCP), cost analysis (breakdowns, trends, and billing insights), monitoring (metrics, alarms, and health checks), security-related checks (vulnerability scanning and compliance visibility), and credential management for securely storing and using cloud credentials. The server is designed to keep credentials securely stored and used only for API calls, with environment-based or MCP-configured credentials supported for flexibility across environments.

Prerequisites:

• Node.js 18+ installed on your system
• npm or yarn
• Access to cloud provider credentials (AWS, Azure, or GCP)
• Git to clone the repository

Steps:

1. Clone the repository:

git clone https://github.com/code-alchemist01/Cloud-mcp_server.git
cd Cloud-mcp_server

2. Install dependencies:

npm install

3. Build the project (if applicable):

npm run build

4. Configure credentials (example for AWS/Azure/GCP) – choose one method per provider:

• AWS: set environment variables or use credentials file as described in the README
• Azure: set AZURE_SUBSCRIPTION_ID and other credentials as needed
• GCP: set GCP_PROJECT_ID and GOOGLE_APPLICATION_CREDENTIALS to your service account key

5. Run the server:

npm start

6. Connect a MCP client by creating an MCP config that points to the running server (see mcp_config for an example).

Tips and common considerations:

• Ensure your cloud provider credentials have appropriate permissions for the actions you intend to perform (e.g., EC2 describe/start/stop, S3 bucket access, Compute/Storage access, etc.).
• If running behind a secure network, consider restricting access to the MCP endpoint and enabling appropriate authentication for clients.
• When using MCP config to supply credentials, prefer environment variables for sensitive data, but you can embed within the MCP config for testing or controlled environments.
• For large-scale operations, monitor SDK rate limits and implement retry/backoff strategies in your client tooling.
• If you encounter issues with credentials loading, verify the GOOGLE_APPLICATION_CREDENTIALS path or the presence of required environment variables in the running process.
• The server is designed to be extended with provider adapters; new providers can be added by following the existing adapter pattern in the repository.