Skip to content
    monk.io
    Get startedStart

    Introducing the universal software operator

    Your agent writes the code.Monk operates your production.

    Across any cloud, any service, any resource — on your own infrastructure.

    Plugin for

    Claude CodeCodexCursor

    Five things your coding agent cannot be. But Monk is.

    1 / 5
    Your coding agent
    task-appMonk plugin connected
    Ask Monk to deploy, change or explain your app…Monk
    Monk · local dashboard
    monktask-app demo

    Feed

    Up to date

    Waiting for your agent.

    The cloud token is typed into Monk's own form, in the browser, not into the chat.

    Day 1. Day 2. Day 200.

    Day 1Ship

    Your agent writes the app. Monk provisions the infrastructure, wires the services together and deploys. You come back to running software.

    Day 2Operate

    Something breaks. Monk finds the cause rather than the symptom and your agent fixes the code — then the repair goes through the same review as any other change.

    Day 200Optimise

    The map is still current. Cost per resource, what-if pricing before you commit, and one command to move a workload to a cheaper cloud.

    Not another tool.The universal software operator itself.

    It understands the system, operates its parts, and carries decisions through to running changes.

    It understands.

    The application. The infrastructure.
    The connections between them.

    Monk keeps a live map of the system it manages: what is running, where it lives, what it depends on and what it costs. That state stays with the application, across agent sessions.

    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.

    It operates everything.

    Every resource has a lifecycle.
    Monk knows how to work with it.

    Read logs, rotate a key, provision a VM or scale a service. Monk connects your agent to the operations each resource supports, across cloud infrastructure and external services.

    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. The selected resources are emphasized as the system changes.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
    API / stdout
    14:32:10INFOGET /tasks200 12ms14:32:11INFOPOST /exports202 8ms14:32:12INFOqueue.publishexport-4214:32:13INFOworker.connectedworker-0314:32:14INFOdb.query18 rows 4ms14:32:15INFOGET /health200 1ms14:32:16INFOauth.verifysession valid14:32:17INFOGET /tasks/42200 6ms14:32:18INFOstorage.putexport-42.csv14:32:19INFOexport.complete1.8s 240KB14:32:20INFOGET /exports/42200 3ms14:32:21INFOqueue.ackexport-4214:32:22INFOPOST /tasks201 9ms14:32:23INFOcache.hittasks:list14:32:24INFOGET /tasks200 2ms14:32:25INFOGET /health200 1ms
    Example stream · read-only
    01 / 05Read the API logs

    Select the API and follow its output without losing its place in the system.

    Read the diagram, step by step

    1. Read the API logs

      Select the API and follow its output without losing its place in the system.

    2. Scale the workers

      One worker becomes three. The container expands around the stack; the dependencies stay attached.

    3. Add a VM

      The same cloud gains a machine. Two of the three workers move onto it, sharing the existing queue and services.

    4. Rotate the Stripe key

      Rotate the credential, then notify the API and every worker that depends on it. The value stays hidden.

    5. Back up the database

      The database stays in place. A new snapshot appears beside it in Atlas.

    AWSGoogle CloudAzureDigitalOceanHetznerMongoDBStripeAuth0

    180+ ready-made integrations. Custom integrations can be built and tested against a service’s API.

    Explore integrations

    It chooses and executes.

    The architecture can change.
    The application keeps its identity.

    Choose where a workload belongs, how much capacity it needs and what the change will cost. Review the plan, approve it, and have Monk carry it out across the system.

    Monk’s application graphVite frontend, Node.js API and Go workers, with shared Redis, Postgres, 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. The selected resources are emphasized as the system changes.DigitalOceanAmsterdam · ams3Amazon RDSAWS · eu-west-1Droplet 01s-2vcpu-4gbDroplet 02s-2vcpu-4gbDroplet 03s-2vcpu-4gbContainerContainerContainerDroplet 04s-2vcpu-4gbContainerFrontendVite · web appAPINode.jsWorkersGo · background jobsDatabasePostgres · managedJob queueRedisAuth0AuthenticationStripePaymentsIngressIngressLoad balancerHTTPS · 443app.example.comCloudflare · DNSAPI volume20 GB · Block storageWeb volume10 GB · Block storageSpacesObject storageDaily backupRDS · daily snapshot
    Drag to explore
    01 / 03DigitalOcean · Amsterdam

    Droplets, Spaces and block storage. The application uses managed Postgres on Amazon RDS.

    Read the diagram, step by step

    1. DigitalOcean · Amsterdam

      Droplets, Spaces and block storage. The application uses managed Postgres on Amazon RDS.

    2. AWS · Frankfurt

      The same application on EC2: t3.medium instances, S3, EBS and an application load balancer. RDS stays connected.

    3. Google Cloud · London

      Compute Engine e2-medium instances, Cloud Storage and persistent disks. Postgres runs on Neon, in the same place in the graph.

    It keeps operating between changes. The orchestrator recovers workloads within configured policies. When an issue needs investigation, Watcher diagnoses it and proposes a repair for your approval.

    You set the limits.
    Monk enforces them.

    Permissions and approval checks live in the orchestrator, below the model. The same controls apply whether a change comes from a person, a pipeline or an agent.

    Your authority.
    The agent acts within your permissions. Access to a resource is checked before it can act.
    Your approval.
    Spending, architecture changes and infrastructure deletion require review. You see the affected resources and estimated cost.
    Your secrets.
    Credentials enter through Monk’s own forms. Workloads receive scoped access; the model uses placeholders.

    Actions are recorded and attributable. Your cloud account stays yours.

    Explore security practices

    Monk, the AI DevOps is here to stay. Monk streamlines the entire system setup and deployment process using AI, turning what used to be a daunting task into something remarkably easy. I'm thoroughly impressed!

    Rajat ShresthaDevOps at KO Technologies

    Just had my mind blown by a monk demo! This AI/ML/LLM thing isn't going to generate a bunch of code…its going to remove it.

    Brian LeRouxCo-founder and CTO at begin.com

    Monk is the big-brained devops friend we've all been waiting for: knowledgeable, tireless, patient, and just opinionated enough.

    Ben WenTallyfor

    Monk is a game-changer so far! It has already saved me hours on project deployment, freeing up more time for development. Looking forward to the full release!

    Mohamed KaramDeveloper at FreeMcServer.Net

    For the app you’re building.
    And the team you’re becoming.

    Building with an agentGet your first app into production.Running a small teamBuild and operate from the same agent.Managing a bigger teamSet access and rules for production.

    Already have a deployment setup? Compare Monk with your current approach.

    Start where
    you are.

    Build a new app, or adopt one you are already running. Either way Monk goes from your repository to a plan you review before anything deploys, and an app that is already serving keeps serving while the first deployment is built alongside it.

    Get started

    Evaluating for a team? See what your ops team controls or talk to us.

    FAQ

    Cloud ownership, approvals and adoption. See pricing.

    My coding agent can already deploy. Why add Monk?

    Coding agents can deploy through commands and tools. Monk adds a maintained graph of the services, infrastructure and connections it manages, so a fresh session can read deployment state and current status. It also supplies deployment approvals, cost tracking for supported resources and workload recovery within configured policies.

    How does Monk charge?

    Monk is moving to pricing by managed capacity. Cloud and coding-agent costs are separate. Capacity plans are a proposal; current subscriptions and checkout have not changed. Confirm terms on the pricing page.

    Where does my app run?

    Resources run in your cloud account. You pay your cloud provider directly, separately from Monk and your coding agent.

    What happens without my approval?

    The orchestrator automatically recovers workloads within configured policies. Watcher adds AI diagnosis; its proposed fixes need your approval. Review the deployment plan and provider costs before provisioning.

    How do I try it with an existing app?

    Monk builds fresh infrastructure from your codebase. Your existing deployment stays untouched. Test the new environment before you switch. Data transfer and traffic cutover are separate steps. Both environments incur cloud costs while they run.