# Northflank MCP for AI Agents AI Agent Connect

> Northflank (Developer Cloud & Orchestration) lets you manage your cloud infrastructure through your AI agent. You can deploy microservices, trigger CI builds, audit background jobs, and manage secret vaults without jumping between tabs. It connects your Northflank account to your agent so you can handle orchestration and deployment via natural conversation.

## Overview
- **Category:** ship-it
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_geMoJ5G8AXHTcf6ry4BH4pzpuf69uz3m4ZSGhSdn/ai-agent-connect
- **Tags:** microservices, cloud-deployment, ci-cd, infrastructure-orchestration, backend-hosting

## Description

Northflank (Developer Cloud & Orchestration) connects your Northflank account to any AI agent so you can handle infrastructure orchestration through natural conversation. Instead of hunting for a specific dashboard to see if a container is running or to check how much memory a service is eating, you just ask your agent. You can check your project metadata, see your secret mappings, or cycle a container to clear out a memory leak, all while staying in the same flow. It handles the heavy lifting of cloud management so you can focus on shipping features. Since Vinkius hosts this, you can get it running in your preferred client and start talking to your infrastructure immediately. It's about moving from manual dashboard navigation to a conversation-based way of managing your cloud lifecycle. This means you can audit your background jobs, trigger fresh builds to refresh production assets, and manage your secret groups without ever leaving your workspace. Whether you are provisioning new project spaces or tearing down entire microservice ecosystems, the agent handles the commands for you. It's a shift from manual configuration to a more fluid way of interacting with your deployment pipeline. This connection lets you verify your environment variable mappings across VPC boundaries and monitor periodically executing processes like heavy database aggregations. You get a direct line to your production environment, allowing you to manage organizational project spaces and verify secret vault mappings across multiple environments efficiently. It removes the friction of jumping between different tools to keep your infrastructure healthy.

## Tools

### create_project
Provision a new isolated Northflank Project space. This sets up the fundamental cloud region routing for your ingress requests.

### list_services
List all explicit application and microservice instances inside a Project. It delineates distinct IDs for both Dockerized and native builds.

### restart_service
Gracefully cycle container replicas for a specific Service. This helps recycle transient RAM accumulation and restore standard execution timings.

### trigger_build
Command Northflank builders to compile and deploy your latest code. Use this to force-refresh production assets or verify recent merges.

### delete_project
Permanently tear down a Project and all cascading microservices. This action severs traffic routing and kills all active cron jobs.

### list_secrets
List Secret Group dictionaries that inject into your environments. This lets you verify environment variable mappings across VPC boundaries.

### get_project
Examine deeper regional metadata for a specific Northflank Project. It helps you correlate the global resource restrictions backing your pipelines.

### get_service
Retrieve the exact structural anatomy of a single Service. This is useful for assessing horizontal scaling needs and verifying SSL associations.

### list_jobs
Enumerate all Batch and Cron Jobs within a Project. You can use this to monitor periodically executing processes like heavy DB aggregations.

### list_projects
List all top-level organizational Projects in your Northflank account. This fetches the mandatory Project IDs you need for management.

## Prompt Examples

**Prompt:** 
```
Are there any cron jobs running in my 'data-processor' project?
```

**Response:** 
```
I've checked the jobs for the 'data-processor' project. Here is what's currently active:

| Job Name | Status | Last Run |
| :--- | :--- | :--- |
| **DB_Aggregation** | Running | 2 mins ago |
| **Security_Scan** | Success | 1 hour ago |
| **Cache_Clear** | Pending | Scheduled |

Would you like me to check the logs for the DB_Aggregation job?
```

**Prompt:** 
```
I need to see the resource limits for the auth-service.
```

**Response:** 
```
I've retrieved the structural anatomy for the **auth-service**: 

*   **CPU Limit**: 1000m
*   **Memory Limit**: 2Gi*   **SSL Status**: Active
*   **Replica Count**: 3

Everything looks standard. Do you need me to adjust any of these boundaries?
```

**Prompt:** 
```
Trigger a new build for the gateway service.
```

**Response:** 
```
Triggering the production build for **gateway** now. 

I've signaled Northflank to pull the latest code from your linked repository. I'll keep an eye on the build cycle and let you know as soon as the new image is live.
```

## Capabilities

### Provision new project spaces
Create isolated project environments with specific regional routing.

### Tear down microservice ecosystems
Permanently remove projects, services, and active cron jobs.

### Trigger fresh CI builds
Command the platform to compile and deploy your latest code.

### Cycle container replicas
Gracefully restart service containers to clear memory accumulation.

### Audit background jobs
List and inspect all active Batch and Cron jobs.

### Inspect secret group mappings
Verify environment variable mappings across VPC boundaries.

### Retrieve service resource limits
Check the exact CPU and RAM boundaries for any service.

## Use Cases

### Fixing a production memory leak
A backend dev notices a service is sluggish. They ask the agent to use restart_service to cycle the replicas and restore performance immediately.

### Verifying a hotfix deployment
A dev just merged a PR. They ask the agent to use trigger_build to force-refresh the production assets without navigating the UI.

### Auditing batch job health
A DevOps engineer needs to see what cron jobs are running. They ask the agent to use list_jobs to check for heavy DB aggregations.

### Bootstrapping a new environment
A platform team needs to set up a new environment. They ask the agent to use create_project to bootstrap the new project space.

## Benefits

- Faster deployments: Use trigger_build to start production builds instantly without opening the Northflank dashboard.
- Easier debugging: Quickly check service anatomy with get_service to see if your app is hitting CPU or RAM limits.
- Better visibility: Audit all your background processes at once using list_jobs to find stuck cron jobs.
- Secure secret management: Verify your .env and DATABASE_URL mappings across VPCs using list_secrets.
- Cleaner project lifecycle: Use create_project and delete_project to manage your infrastructure as code via your agent.
- Improved reliability: Use restart_service to quickly recycle containers and fix transient memory issues in production.

## How It Works

The bottom line is you get a conversational interface for your entire Northflank cloud infrastructure.

1. Subscribe to the Northflank MCP on Vinkius.
2. Enter your Northflank API Token into your AI client.
3. Ask your agent to list projects, check service health, or start a build.

## Frequently Asked Questions

**Can I use Northflank MCP to manage my microservices?**
Yes, you can manage your microservice instances, check their resource limits, and handle deployments directly through your AI agent.

**How do I see my secret groups with Northflank MCP?**
You can ask your agent to list secret group dictionaries. This helps you verify your environment variable mappings across different VPC boundaries.

**Can I restart a container using Northflank MCP?**
Yes, you can use the agent to gracefully cycle container replicas. This is a quick way to clear out transient memory issues and restore standard execution timings.

**Does Northflank MCP support triggering builds?**
Yes, it can signal Northflank to pull your latest code and start a new build cycle. This is great for forcing a refresh on production assets.

**Can I use Northflank MCP to provision new projects?**
Yes, you can use the agent to create new isolated project spaces or permanently tear down existing ones, helping you manage your cloud lifecycle.

**Is Northflank MCP good for auditing background jobs?**
Yes, it allows you to enumerate all Batch and Cron jobs within a project. You can easily monitor periodically executing processes like heavy database aggregations.

**Can I restart my application containers through the agent without a full rebuild?**
Yes. Use the `restart_service` tool by providing your Project ID and Service ID. Your agent will signal Northflank to gracefully cycle the container replicas, effectively refreshing the running environment while maintaining your current code version.

**How do I trigger a fresh build and deployment from my GitHub repository?**
The `trigger_build` tool hooks into your linked VCS. Your agent will command Northflank to fetch the latest commits from your repository and start an automated build and release cycle, promoting the new image to production immediately.

**Can my agent check the status of my periodic cron-jobs?**
Absolutely. Use the `list_jobs` tool to identify all isolated background processes within a project. Your agent will report the current execution statuses and schedule parameters for both ad-hoc batch tasks and recurring cron-jobs.