✦ MCP-native container ops
Point your repo, push your code, app goes live.
Your AI can manage it too.
DockPod is a single-binary container ops panel + lite PaaS. Deploy from a pasted compose file or a git repo, provision databases, and let Claude or Cursor run the whole thing from a conversation.
curl -fsSL https://dockpod.io/install.sh | sh
One panel, two workflows
Already running containers on a VPS
You don't want to pay for another server, or install five containers just to get a management UI.
- Full container/image/volume/network control
- Live logs and live stats, no extra agent
- Docker or Podman — same binary, same UI
- Scheduled maintenance: prune, restart, auto-update
You generate code with AI, now you need it live
Claude or Cursor just wrote your app and a docker-compose.yml. You don't want to think about infrastructure.
- Push to GitHub, point DockPod at the repo
- Auto-detects compose vs Dockerfile
- Webhook auto-deploy on every push
- Tell Claude "deploy my app" — it just works
Your AI already knows how to deploy it
DockPod runs a full Model Context Protocol server — Claude, Cursor, or any MCP client can inspect and manage containers directly from the conversation.
"Build me a Laravel app with PostgreSQL" → your AI generates code + docker-compose.yml
Push to GitHub → point DockPod at the repo → webhook auto-configured
Tell Claude: "deploy my app" → MCP triggers the deploy, streams the logs, done
One config block, no server-side setup — works with Claude Desktop, Cursor, or Claude Code
"mcpServers": {
"DockPod": {
"command": "npx",
"args": ["mcp-remote", "http://your-server:8090/sse", "--header", "Authorization:${AUTH_HEADER}"],
"env": { "AUTH_HEADER": "Bearer dp_mcp_xxxxxxxxxxxx" }
}
}
}
Container ops + lite PaaS, in one binary
Container Management
Start, stop, restart, remove. Live logs and live stats over WebSocket. Auto-refreshing lists, no reload needed.
Paste & Deploy
Paste a docker-compose.yml, or pick from common presets — databases, MinIO, RabbitMQ, Meilisearch, Grafana, and more. No git required — free, forever, unlimited.
Git Deployment
Point at a repo. Compose, Dockerfile, or zero-config Nixpacks/Railpack auto-detected. Builds on every push via webhook, with live streaming logs.
Custom Domains
Point a real domain at any deployed app — DockPod manages a shared Traefik instance and issues Let's Encrypt certificates automatically.
Database Provisioning
One-click Postgres, MySQL, Redis, MongoDB, ClickHouse, or DragonflyDB — each in its own isolated, localhost-bound container. Backup on demand.
Git Providers
Connect GitHub in one click — DockPod creates its own App, no tokens to copy. GitLab, Bitbucket, and Gitea connect with a personal access token.
Built-in Security
Per-user two-factor authentication, a full audit trail, and reverse-proxy network protections — every MCP tool call attributed to the key that made it.
Notifications
Email, Telegram, Slack, or Discord — get pinged the moment a deploy succeeds or fails.
Scheduled Tasks
Cron-based image/volume pruning, container restarts, and watchtower-style auto-update.
Multi-user & Roles
Invite your team with admin/member roles. Full audit log of every account-changing action.
MCP-native
A real Model Context Protocol server, not a bolt-on. Free tier gets read-only tools; Pro gets full action tools.
Live Resource Tuning
Adjust memory, CPU, and process limits on a running app — or any container — with no rebuild and no restart.
Questions, answered
The short version of what DockPod is and isn't.
What is DockPod?
A self-hosted container ops panel and lite PaaS — a single Go binary you install on your own VPS. It gives you app deployment, container management, and database provisioning without handing your infrastructure to a third-party platform.
How is app deployment different from Railway or Render?
DockPod runs on hardware you already own or rent — no per-request billing, no vendor lock-in. You get the same "push your repo, app goes live" workflow, just self-hosted.
Does it work with AI coding tools?
Yes — DockPod runs a real Model Context Protocol (MCP) server. Claude, Cursor, or any MCP client can inspect containers, trigger a deploy, and read back logs directly from the conversation, which is exactly what a vibe coder workflow needs.
What can I deploy?
Paste a docker-compose file for a one-off deploy, or connect a git repo and let DockPod build it with a Dockerfile, Nixpacks, Railpack, or its own compose file on every push.
Is DockPod free?
Yes, the Free tier is free forever — full container management, unlimited Paste & Deploy, 3 Git Deployment projects, and 3 databases. Pro is $49/year if you outgrow those limits.
What do I need to install it?
A Linux VPS (Ubuntu, Debian, or an RHEL-family distro) and Docker or Podman — the install script sets up whatever's missing and gets the panel running in one command.
Free forever. Pro when you need it.
No credit card, no trial countdown. Upgrade when you actually hit a limit.
Free
- Full container/image/volume management
- Paste & Deploy — unlimited
- Git Deployment — 3 projects, 5 latest builds of history
- Database provisioning — 3 databases
- Email notifications
- Full MCP read-only tool access
Pro
- Unlimited Git Deployment projects + history
- Rollback to any previous build
- Unlimited database provisioning (Postgres/MySQL/Redis/Mongo/ClickHouse/Dragonfly)
- Telegram, Slack & Discord notifications
- Scheduled tasks + auto-update
- Multi-user, roles, and audit log
- Full MCP action-tool access