Skip to content

    Deploy your app in your cloud.

    Take the frontend, API and supporting services from your repository to your cloud. Review the plan in Monk, then deploy from your coding agent.

    One application, every dependency.

    Your app is more than a frontend. Monk maps its services and dependencies, prepares the infrastructure and connects the pieces. You review the plan before it creates resources.

    Monk’s application graphVite frontend, Node.js API and Go workers, with shared Redis, MongoDB, Stripe, Auth0 and object storage dependencies. Services are contained in containers and cloud machines, with frontend and API ingress, load balancing, DNS, volumes and backups. Workers connect to the queue directly. Arrows show connections; moving dashes show example data flow. DigitalOceanAmsterdam · your accountMongoDB AtlasManaged dataDroplet 012 vCPU · 4 GBDroplet 022 vCPU · 4 GBDroplet 032 vCPU · 4 GBContainerContainerContainerDroplet 04ContainerFrontendVite · web appAPINode.jsWorkersGo · background jobsDatabaseMongoDBJob queueRedisAuth0AuthenticationStripePaymentsIngressIngressLoad balancerHTTPS · 443app.example.comCloudflare · DNSAPI volume20 GB · block storageWeb volume10 GB · block storageObject storageDigitalOcean SpacesDaily backupAtlas · daily snapshot
    Drag to explore
    08 / 08The complete operating model

    Volumes, shared object storage and Atlas backups. One live graph, down to the resources underneath.

    Read the diagram, step by step
    1. The application

      A frontend, an API, workers and a database. Start with what the application does.

    2. The dependencies

      The API and workers both use the database. The frontend talks to the API.

    3. The work between services

      A Redis queue connects the API to its workers. Jobs have somewhere to wait.

    4. The outside world

      API and workers use Stripe. Frontend and API share Auth0. These are part of the system too.

    5. Where it actually runs

      DigitalOcean droplets sit beneath the application. MongoDB lives in Atlas.

    6. Inside each machine

      Containers appear inside the droplets, around the services they run.

    7. How traffic gets there

      A Cloudflare domain points to the load balancer, which routes to the frontend and API ingress. Workers consume jobs from the queue.

    8. The complete operating model

      Volumes, shared object storage and Atlas backups. One live graph, down to the resources underneath.

    From repository to running app.

    1. Start from the repository

      Open your app in a supported coding agent with Monk installed. Ask for a deployment plan. Monk identifies the app components and dependencies, then asks for missing settings and credentials.

    2. Review the system it will create

      Inspect the proposed resources, service connections, warnings and estimated cloud costs in the local approval screen. Resolve unpriced resources and unsupported dependencies before proceeding.

    3. Deploy and check a real task

      Approve the plan, follow deployment progress and open the app. Exercise a workflow that touches its dependencies; a loaded page alone does not verify the whole application.

    Check the complete workflow.

    Verify a real application task across the frontend, API, database and workers. The linked task/export example demonstrates this with a checked-in infrastructure manifest; review and adapt it for your application.

    What if I already have a deployment?

    Monk builds fresh infrastructure from your codebase. Your existing deployment stays untouched. Test the new environment before you switch. Moving production data and switching traffic are separate steps.

    What do I need before I start?

    A supported coding agent, the Monk plugin, an app repository, a Monk account and access to the target cloud. Provider credentials are entered through the companion’s credential form, not pasted into chat.

    What if a dependency is unsupported?

    Ask Monk to identify unsupported services during analysis. Review the alternative configuration or contact support before approving resources. Do not assume every service has an equivalent on every cloud.

    Try it with your application.

    Start from your repository. Review the resources and costs before you deploy.

    Bring an existing app Start a new app