MCTL

First-user checklist ​

Use this checklist immediately after your MCTL tenant is provisioned. It covers every access and setup step required before you run your first workload. Each item links to the relevant deep-dive guide or URL for more detail.

  • [ ] 1. Confirm your tenant Your tenant is the Kubernetes namespace that owns your services, resource quotas, and access controls. Sign in at app.mctl.ai and verify that your tenant appears on the dashboard. If you need to create one, follow the Tenants guide.

  • [ ] 2. Verify Portal access Open app.mctl.ai in a browser and confirm that you can sign in and see your tenant's dashboard. If you encounter a login error or missing tenant, contact your platform operator before proceeding.

  • [ ] 3. Verify ArgoCD access Open ops.mctl.ai and confirm that you can sign in and see your tenant's applications. ArgoCD is the GitOps engine that manages continuous deployment for your services.

  • [ ] 4. Verify Argo Workflows access Open workflows.mctl.ai and confirm that you can sign in and view the workflow namespace for your tenant. Argo Workflows runs the build and deployment pipeline submitted by mctl_deploy_service in Step 6 below. Auto-deploy on every push to your repository is optional and requires the CI job described in the Scaffolding guide — it is not enabled by default.

  • [ ] 5. Prepare a Git repository or Docker image Your workload must be reachable from MCTL. For source-based deployments, push your application code to a GitHub repository and ensure you have a Dockerfile (or use MCTL scaffolding to generate one). See Deploy your first app - Step 2 and the Scaffolding guide for details. If you already have a container image published to a registry, skip straight to Deploy a Service instead — it deploys from an existing image reference with no repository or Dockerfile required.

  • [ ] 6. Deploy your first workload Follow the Deploy your first app guide to connect your repository to MCTL, run the onboarding workflow, and obtain a live service URL. The MCP path is fully documented end to end; the portal path currently covers tenant and dashboard steps only — repository access and sync are MCP-only for now (see the tips in Deploy your first app).

  • [ ] 7. Review logs and deployment status After deployment completes, confirm the service is healthy by checking the workflow run in Argo Workflows and reviewing application logs. You can also inspect status and retrieve logs through the MCP Tools Reference if you have an AI client connected.

  • [ ] 8. Manage your platform via MCP MCTL exposes its full management surface through a Model Context Protocol server, allowing you to deploy, scale, and inspect services from any MCP-compatible AI client. Start with the MCP Overview, then follow the Connecting guide to wire up your client.