digitalocean

This MCP server provides programmatic access to DigitalOcean's API through a comprehensive set of 48 tools spanning droplets, volumes, snapshots, images, floating IPs, load balancers, firewalls, Kubernetes clusters, and container registries. It exposes a consistent MCP interface for managing infrastructure, allowing you to perform common operations such as listing resources, creating new instances, updating configurations, attaching storage, and testing connectivity. Use the provided tools to automate infrastructure workflows and integrate DigitalOcean management into your own MCP-enabled clients.

To interact with the server, you will use the standardized MCP client protocol. The server supports a broad suite of tools grouped by service category. For example, you can list droplets with list_droplets, create a new droplet with create_droplet, manage volumes via create_volume or attach_volume, configure load balancers with create_load_balancer, and enforce security with firewall-related tools. The toolkit also includes test_connection to verify API access and authentication before performing other operations. Each tool accepts a specific set of arguments, so refer to the examples in the README to structure your requests correctly and handle responses in your client application.

Prerequisites:

• Docker installed and running on your host
• A DigitalOcean API token with appropriate scopes

Option 1: Run via Docker (recommended for quick start)

1. Ensure you have a valid DigitalOcean API token prepared: export DIGITALOCEAN_ACCESS_TOKEN="your_digitalocean_api_token"
2. Start the MCP server container: docker run -i -e DIGITALOCEAN_ACCESS_TOKEN="$DIGITALOCEAN_ACCESS_TOKEN" ghcr.io/rohit-kaundal/digitalocean-mcp-server:latest
3. The server will expose its MCP interface via stdio for the client to connect.

Option 2: Build and run locally from source (Go)

1. Prerequisites: Go 1.19+ and a configured GOPATH/module environment.
2. Clone the repository: git clone https://github.com/rohit-kaundal/digitalocean-mcp-server.git
3. Build the binary: cd digitalocean-mcp-server go mod tidy go build -o digitalocean-mcp-server
4. Run the server (with token set in environment): export DIGITALOCEAN_ACCESS_TOKEN="your_digitalocean_api_token" ./digitalocean-mcp-server
5. The server will start and listen for MCP requests via stdio transport.

Environment Variables:

• DIGITALOCEAN_ACCESS_TOKEN: Required. Your DigitalOcean API token. The server will not operate without this token.

Common issues:

• Ensure the token has the necessary scopes for the managed resources.
• If using Docker, ensure the container has access to your environment variables or token file as needed.
• When running locally, keep the token secret and avoid exposing it in logs or version control.

Configuration options:

• You can extend or override tool arguments per client request by adjusting the MCP client payloads to match the tool names and expected parameters documented in the README.