Dockpeek is a small self-hosted dashboard for the containers already running on a Docker host. Instead of remembering which host publishes which port, you get one table of every container with its published ports, image and health, and a click on a port opens that service’s web interface in a new tab. It is the practical answer to a homelab that has grown to a dozen compose stacks whose dashboards live on scattered IP:port addresses.
The app itself is a Flask application served by gunicorn inside a container, and the only thing it needs from the host is the Docker socket (read-write for updates, or a socket proxy if you prefer to limit access). Configuration is a handful of environment variables – a SECRET_KEY and one username/password pair for the dashboard login – so there is no database to back up and no first-run wizard.
DOCKER_HOST_n_URL variables; remote hosts only need the Docker API exposed, not Dockpeek itself.dockpeek.tags groups containers, dockpeek.link turns a name into a link, dockpeek.https forces HTTPS on a port and dockpeek.ports adds ports that Docker does not publish.Dockpeek sits alongside tools such as Portainer, Yacht and Homepage, but deliberately does less: it reads the Docker API and shows what is exposed rather than trying to be a full container management console. If the need is to deploy stacks, manage volumes or handle RBAC, a heavier control panel is the right tool; if the need is a fast index of what is running on the box and how to reach it, this covers it.
A single docker run or compose service is enough: mount /var/run/docker.sock, set SECRET_KEY, USERNAME and PASSWORD, and publish the web port (the documentation uses 3420). For anything exposed beyond localhost, the project documents a socket-proxy deployment so the app never holds the full Docker socket, and authentication is a single shared login rather than per-user accounts – it belongs behind a reverse proxy or on a private network.
Open source, MIT licensed. Self-hosted and free to use, modify and redistribute; the source is on GitHub and images are published on Docker Hub and ghcr.io.