podman

The Podman MCP Server provides a modular, AI-assisted MCP (Model Context Protocol) server that can interface with Podman (and Docker) to manage containers, images, and related operations. It exposes a CLI with HTTP/streaming endpoints and a set of tooling commands under the Tools section, allowing you to inspect containers, list resources, fetch logs, remove containers, and run new containers via a consistent MCP interface. You can run the server in HTTP mode for real-time communication and SSE updates, or in STDIO mode for direct integration with other tooling. Use the available transport modes to tailor how you send commands and receive results from your MCP workflow. The server is designed to auto-detect Podman backends (api via Unix socket or cli) and provides a way to force a backend if needed.

Prerequisites:

• Node.js and npm (for npx usage)
• Podman (or Docker) installed on the host

Installation steps:

1. Ensure Node.js and npm are installed on your system.
2. Install and run the Podman MCP Server using npx:

# Quick start with npx (no global install required)
npx podman-mcp-server@latest --help

3. If you prefer, you can download and run the binary directly (refer to the repository's latest release).
4. Optionally, verify Podman is accessible (via api socket or CLI) so the server can auto-detect the best backend.

Tips and common considerations:

• The server supports multiple transport modes; use --port to enable HTTP and SSE endpoints (e.g., /mcp and /sse).
• Podman backend selection: the server auto-detects the best available implementation (api via Unix socket preferred, cli as fallback). You can force a backend with --podman-impl (e.g., --podman-impl=api or --podman-impl=cli).
• If you run behind a firewall or reverse proxy, ensure the HTTP port you expose is reachable and SSE endpoints are allowed.
• The Tools section includes commands for container_inspect, container_list, container_logs, container_remove, container_run, and container_stop to manage containers programmatically via MCP.
• Environment variables may be needed for advanced configurations (not required for a basic run); refer to the documentation for environment variable placeholders and descriptions.