kubefwd

kubefwd is a Kubernetes-focused port-forwarding utility that bulk forwards all services in a namespace to your local machine, assigning each service a unique loopback IP address. This enables you to interact with cluster services as if they were running locally, using the service names (for example http://api-service:8080 or mysql -h database). The MCP integration adds AI-assisted automation and control capabilities, allowing you to query or control port-forwarding behavior via an API, enabling workflows driven by an AI assistant. Use the interactive TUI mode to monitor traffic, view per-service metrics, and reconnect automatically when pods restart. You can also run kubefwd via Docker, providing a portable way to forward services without installing binaries locally.

Prerequisites:

• Docker installed and running (for the Docker-based usage shown in the MCP config).
• kubectl configured to access your Kubernetes cluster.
• Optional: root/sudo access if using host network changes or editing /etc/hosts.

Installation options:

Option A: Docker (recommended for MCP integration)

1. Install Docker on your machine.
2. Run kubefwd via Docker (as configured in this MCP setup):

# From the MCP manifest above, the server is wired to run the Docker image txn2/kubefwd with the following command:
# docker run -i txn2/kubefwd services -n default --tui

Option B: Prebuilt binaries (local if you prefer to install a binary directly)

1. Visit the kubefwd releases page and download the appropriate binary for your OS.
2. Place the binary in a directory on your PATH, e.g. /usr/local/bin.
3. Run: kubefwd svc -n <namespace> --tui

Option C: Homebrew (macOS)

1. brew install kubefwd
2. Run: kubefwd svc -n <namespace> --tui

Note: If you’re using MCP integration, ensure the environment is configured to allow REST API access to kubefwd’s MCP endpoints as documented by the project.

Tips:

• The unique IP per service (127.x.x.x) means you usually won’t encounter port conflicts when multiple services are forwarded.
• kubefwd updates /etc/hosts to resolve service names as in-cluster DNS names; ensure your /etc/hosts is writable by the process (use sudo if needed).
• If pods restart or connections drop, kubefwd’s auto-reconnect will re-establish forwards automatically.
• When using the MCP integration, you can manage forwards and retrieve status through the provided REST API endpoints; consult the MCP integration documentation for commands and auth requirements.
• For Windows users, consider using the Docker approach or Windows-compatible binaries to run the tool.
• If you run into permission or network issues, verify that required ports are not blocked by your firewall and that your kubeconfig allows API access.