Skip to main content
Getting an app from a repository to running infrastructure usually means Dockerfiles, cloud setup, config and a lot of switching between tools. Monk does it from the coding agent you already use. Your agent reads the code and writes the files. Monk plans, builds, deploys and keeps track of what is running. Everything runs in your own cloud account.

​
What you can do

​
Plan a deployment from the repository

Monk reads your repository and works out the services, what they connect to and what configuration is missing. The plan opens in the local dashboard with the cloud, region, machine sizes and an estimated cost. Nothing is created until you approve it. How: Ask Monk. You approve the plan in the local dashboard. See Deploy an existing app.

​
Start a new app or bring an existing one

For a new app, your coding agent writes the code and Monk plans the infrastructure for it. For an app that already runs somewhere else, Monk builds fresh infrastructure from your source alongside the original and leaves the original alone. Moving data and switching traffic are separate steps you review. How: Ask Monk. See Start a new app and Deploy an existing app.

​
Build and deploy

Monk builds your images at deploy time, for the CPU architecture of the target cluster. An ARM laptop deploying to an x86 cluster works without extra steps. Your coding agent never builds images itself. After the deploy, Monk keeps a model of the running app, so later requests about logs, costs or scaling start from what is actually running. How: Ask Monk. You approve the plan in the local dashboard. See Start a new app.

​
Redeploy as the code changes

Ask Monk to deploy again and it rebuilds the images that changed. Every redeploy goes through the same plan review. The plan says which services are updated by starting the new copy first and which are replaced, which means a brief gap. How: Ask Monk. You approve the plan in the local dashboard. See Talking to Monk.

​
Keep the work split clear

Your coding agent writes the Dockerfiles and Monk’s MANIFEST, and you can commit both with your code. Monk validates them and does not edit your application code. When a deploy needs a code change, your coding agent makes it. How: In your Monk config, written by your coding agent. See How Monk works.

​
Try it

​
Good to know

  • Deploys need approval. Credentials go through forms in the local dashboard, never through chat.
  • Cloud charges are billed by your provider. Monk’s estimate shows the machines and services in the plan.
  • Rolling back means redeploying an earlier commit. Monk does not roll back automatically when health checks fail, and it does not roll back data.
  • A new environment starts empty. Production data is not copied.

Deploy an existing app

Fresh infrastructure next to what already runs

Start a new app

From first commit to a running app

Talking to Monk

What to ask and what you approve

Deployments

The MANIFEST, plans and redeploys

How Monk works

Who writes what

First deployment

A walkthrough