Northflank MCP. Manage microservices, deployments, and cloud resources via chat.

You've got full control over Northflank’s cloud infrastructure using your AI agent. This isn't just a dashboard wrapper; it gives you direct command access across the entire deployment lifecycle—all through natural conversation. Forget clicking around menus; you tell your agent what to do, and it executes complex actions against core Northflank systems.

Project Scope Management: You can start fresh or wipe out environments completely. Use create_project when you need a brand-new, isolated container for resource management. To scope out where everything lives, run list_projects to fetch every top-level organizational project ID. If you need deep details on that boundary—like checking regional resource restrictions—you’ll use get_project.

When the whole thing is garbage and needs going, running delete_project permanently tears down the entire project boundary and all attached microservices.

Service Auditing & Lifecycle: You can't manage what you can't see. To get a full manifest of every active application instance inside your project, call list_services. But sometimes you need to drill into one specific service's DNA. get_service gives you the exact structural anatomy, letting you check its precise CPU throttling limits and RAM allocation boundaries.

If things slow down or memory gets muddy, you don’t panic; running restart_service gracefully cycles container replicas, clearing transient memory and resetting normal execution timings.

Deployment & Automation: When code changes, you need action. Send a command to the Northflank builders using trigger_build to compile and deploy the latest code pulled straight from your linked version control systems. If you've got scheduled tasks—batch jobs or cron routines—you can list them all with list_jobs and inspect exactly what’s running inside that project boundary.

Security & Metadata: Keeping secrets safe is job one. You don't want to manually check every environment variable, so use list_secrets. This tool lists the entire secret group dictionary, showing you which environment variable mappings—like database URLs or API keys—are available to your services and jobs across different VPC boundaries.

To keep tabs on what resources are running in general, you’ll also want list_services inside a project boundary.

This suite allows your agent to manage everything from initializing an isolated development space using create_project, auditing the resource limits of any single service with get_service, and then deploying the fixes by triggering a build. You can list all necessary components—from top-level organizational projects via list_projects to every background job running through list_jobs.

It gives you an end-to-end pipeline: you define the scope, you inspect the resources, you run the jobs, and you deploy the changes.