This server provides an MCP interface to manage Podman containers, images, pods, networks, volumes, and system information for both rootful and rootless Podman installations.
List, inspect, start, stop, restart, kill, remove, and prune containers.
Retrieve container logs, process lists, and resource usage statistics.
Create new containers with custom images, commands, environment variables, labels, and volume mounts.
List, inspect, pull, remove, and prune images.
List, inspect, start, stop, restart, remove, and create pods.
List and inspect networks.
List and inspect volumes.
Get Podman system information and disk usage.
Provides tools for managing Podman containers, images, pods, networks, volumes, and system information via the Podman REST API.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-podman-crunchtools list all running containers"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server for Podman container management via the Podman REST API. Manages containers, images, pods, networks, volumes, and system info. Supports both rootful and rootless Podman.
# uvx (zero-install)
uvx mcp-podman-crunchtools
# pip
pip install mcp-podman-crunchtools
# Container
podman run -v /run/podman/podman.sock:/run/podman/podman.sock:z \
--security-opt label=type:container_runtime_t \
quay.io/crunchtools/mcp-podmanRelated MCP server: podman-mcp
Variable | Required | Default | Description |
| No | Auto-detect | Unix socket path |
| No | — | File containing socket path |
| No | 30 | Request timeout in seconds |
Socket auto-detection order:
$XDG_RUNTIME_DIR/podman/podman.sock (rootless)
/run/user/$UID/podman/podman.sock (rootless fallback)
/run/podman/podman.sock (rootful)
claude mcp add mcp-podman-crunchtools \
--env PODMAN_SOCKET=/run/podman/podman.sock \
-- uvx mcp-podman-crunchtoolsTool | Description |
| List containers |
| Get container details; environment values are redacted, names kept |
| Start a container |
| Stop a container |
| Restart a container |
| Send signal to container |
| Remove a container |
| Get container logs |
| List processes |
| Resource usage |
| Create a container |
| Remove stopped containers |
Tool | Description |
| List images |
| Get image details |
| Pull from registry |
| Remove an image |
| Remove unused images, with the options of |
Tool | Description |
| List pods |
| Get pod details |
| Start a pod |
| Stop a pod |
| Restart a pod |
| Remove a pod |
| Create a pod |
Tool | Description |
| List networks |
| Get network details |
Tool | Description |
| List volumes |
| Get volume details |
Tool | Description |
| Podman system info |
| Disk usage |
AGPL-3.0-or-later
Create a new container.
Inputs: image (required), env, name, labels, +2 more. Returns: object
Get detailed information about a container.
Inputs: name (required). Returns: object
Send a signal to a container.
Inputs: name (required), signal. Returns: object
List containers.
Inputs: limit, filters, all_containers. Returns: object
Get container logs.
Inputs: name (required), tail, since, timestamps. Returns: object
Remove all stopped containers.
Inputs: no named parameters. Returns: object
Restart a container.
Inputs: name (required), timeout. Returns: object
Remove a container.
Inputs: name (required), force, volumes. Returns: object
Start a stopped container.
Inputs: name (required). Returns: object
Get container resource usage statistics.
Inputs: name (required). Returns: object
Stop a running container.
Inputs: name (required), timeout. Returns: object
List processes running inside a container.
Inputs: name (required), ps_args. Returns: object
Get detailed information about an image.
Inputs: name (required). Returns: object
List images.
Inputs: filters. Returns: object
Remove unused images, like podman image prune.
With no arguments only dangling (untagged, unused) images are removed.
Inputs: all, filters, external, build_cache. Returns: object
Pull an image from a registry.
Inputs: reference (required). Returns: object
Remove an image.
Inputs: name (required), force. Returns: object
Get detailed information about a network.
Inputs: name (required). Returns: object
List networks.
Inputs: filters. Returns: object
Create a new pod.
Inputs: name (required), infra, share, labels. Returns: object
Get detailed information about a pod.
Inputs: name (required). Returns: object
List pods.
Inputs: filters. Returns: object
Restart a pod and all its containers.
Inputs: name (required). Returns: object
Remove a pod and all its containers.
Inputs: name (required), force. Returns: object
Start a pod and all its containers.
Inputs: name (required). Returns: object
Stop a pod and all its containers.
Inputs: name (required), timeout. Returns: object
Get Podman disk usage.
Inputs: no named parameters. Returns: object
Get Podman system information.
Inputs: no named parameters. Returns: object
Get detailed information about a volume.
Inputs: name (required). Returns: object
List volumes.
Inputs: filters. Returns: object
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 30 tools
Each tool targets a distinct resource+action pair (container_start vs container_stop vs container_kill), and the resource prefix separates container/image/pod/network/volume tools cleanly. Minor potential confusion between container_kill and container_stop, and between prune vs rm, but descriptions clarify these.
Every tool follows the same resource_action_tool pattern (container_list_tool, image_pull_tool, pod_restart_tool), making the surface highly predictable. No mixed conventions.
30 tools is above the comfortable band, but the breadth is justified by six distinct resource domains (container, image, pod, network, volume, system). Each tool earns its place, though the set feels heavy for a single server.
Containers, images, and pods have solid lifecycle coverage (list/inspect/create/start/stop/restart/rm/prune). However, networks and volumes only have list/inspect with no create/rm, and container exec, image build/push, and container run are missing, leaving notable gaps.
Start, pause, resume, and run commands and file operations in Linux sandboxes.
Linux server administration over OAuth MCP: shell, Docker, systemd, files and monitoring.
Provides capabilities that let LLM agents perform a range of infrastructure management tasks.
Create, watch, pay for and connect hosted AI agent pods on AgentsPodium. Needs an API key.