What is Monk?
Monk is the operations agent for your app, used from your coding agent. It plans infrastructure and service connections, deploys to your own cloud account after you approve, and keeps a model of what it runs between sessions. See Overview.Which coding agents does Monk work with?
The installer sets up Claude Code, Codex, Cursor, Devin, Google Antigravity, VS Code (Copilot), GitHub Copilot CLI, Gemini CLI, OpenCode, Kilo Code, Goose, Zed, Cline, Roo Code, Amp, Warp and JetBrains Junie. Claude Code, Codex and Cursor are the main supported hosts. Any other agent that supports MCP servers over HTTP can connect tohttp://127.0.0.1:7419/mcp. See Agents.
Is it safe to run the installer with curl | sh?
The installer shows what it will change and asks before changing anything, and it checks every download against a list signed with Monk’s release key. Add --dry-run to see the plan without changing anything. See Is it safe to pipe into a shell?
Which clouds can I deploy to?
AWS, Google Cloud, Azure, DigitalOcean and Hetzner, using your own account. You can also deploy to your own machine. See Credentials.Can I use Monk with an app that’s already running?
Yes. Monk builds a new environment from your source code alongside the current one and leaves the original alone. It doesn’t copy your data, undocumented settings or external account state. Moving data and switching traffic are separate steps you review, and both environments cost money while they run. See Deploy an app that’s already running.Do I need to write Dockerfiles or Kubernetes configuration?
No. Your coding agent writes the Dockerfiles and Monk’sMANIFEST, and Monk validates them before it plans. Monk runs workloads on its own orchestrator, not Kubernetes. See How Monk works.
Where do my credentials go?
Into a form in the local dashboard on your computer, never into the chat. Monk keeps them in your OS keychain (or an encrypted file where there is none) and pushes them to your cluster when a deployment needs them. See Security.What does Monk do without asking me?
It doesn’t create infrastructure or deploy until you approve the plan, which shows the services, connections and estimated cost. Deletions and other privileged actions also wait for your approval in the local dashboard. Within an approved deployment, the orchestrator restarts and reschedules managed workloads under its policies.What happens when something breaks?
The orchestrator recovers managed workloads within its policies and retry limits. Watcher, if you set it up, sends crash and resource alerts with an AI-written diagnosis to Slack. Any fix Watcher proposes needs a person’s approval. See Watcher and alerts.Can I control Monk from Slack?
No. Slack only receives Watcher alerts. You can’t chat with Monk or approve anything from Slack.Does Monk set up CI/CD?
Yes, with GitHub Actions. You provide a GitHub token through the local form and review the setup before Monk creates the workflow. See CI/CD.Does Monk work for teams?
Yes. Organizations have people, roles and permissions, and the Activity log in the web dashboard records what happened. Each project can have its own environments.How much does Monk cost?
New accounts get a 5-day trial with every Pro feature and no card. Sign up at monk.io, and see monk.io/pricing for plans. Your cloud provider bills you for the resources Monk creates, and your coding agent’s subscription is separate. See Costs for estimates and billing alerts.How do I update Monk?
Run the install command again. It updates Monk and the plugin in every agent it set up. See Updating.How do I report a bug or ask for a feature?
Ask your agent/monk report a bug, use Report a bug or an idea in the web dashboard, or post on the community forum. See Getting help.