<!-- Generated from the canonical HTML page. Do not edit separately. -->
# How Ankra Works - From Repository to Running Application

Canonical: https://ankra.ai/how-it-works

Connect a repository and a Kubernetes cluster. Ankra writes the setup, tests, builds and scans every push in your cluster, deploys it and gives you a live URL.

How it works

# Ankra gives you a path from your repository to your  running system.

Connect a repository and a Kubernetes cluster. Ankra writes the setup, tests, builds and scans every push in your own cluster, deploys it and gives you a live URL. Then it helps you keep it running.

[Get your app live](https://platform.ankra.app/auth/signup)  [Try a playground](https://ankra.ai/playground)

30 vCPU of platform usage included. Try it without a card; a card unlocks full access and €59 of credit.

ankra application ship . --cluster production

Repository analysed

Setup pull request opened

Tests passed

Image built in your cluster

Scanned with Semgrep, Checkov and Trivy

Published sha-4f2c9a1

Deployed, 2 pods running

Live:  https://shop.example.com

The same flow runs from the portal, the CLI or your AI agent.

From repository to live

## Four steps.  Most of them are ours.

You connect, review and push. Ankra does the container, pipeline, scanning and deployment work in between.

1. 01

  ### Connect your code and a cluster

  You

  Connect a GitHub repository. Import an existing Kubernetes cluster with one Helm command, create one on your cloud, or start in a playground.

  Ankra

  Links the cluster through an outbound-connected agent, so there is no inbound port to open.

  You get `A cluster ready for your app`

   [Choose your infrastructure](https://ankra.ai/features#clusters)
2. 02

  ### Review the setup pull request

  You

  Read the pull request and merge it like any other change.

  Ankra

  Reads your code and writes the container, Kubernetes manifests and pipeline it needs.

  You get `Dockerfile``.ankra/pipeline.yaml``.ankra/manifests/`

   [Explore applications](https://ankra.ai/features#applications)
3. 03

  ### Push your code

  You

  Keep working the way you already do. Push to your branch.

  Ankra

  Tests, builds and scans every push inside your own cluster. Only an image that passes the gate is published.

  You get `A scanned image``An immutable sha- tag`

   [Explore delivery](https://ankra.ai/features#pipelines)
4. 04

  ### Go live

  You

  Open the link and share it.

  Ankra

  Deploys the new image, checks that pods are running and gives you the address.

  You get `Live: https://…`

   [Explore operations](https://ankra.ai/features#operate)

See it on real infrastructure

## Follow a change from the repository to production.

One flow, run three ways in the console, the CLI and Claude Code: a pull request, a security gate that stops a critical vulnerability, the fix, staging, and a production deploy that waits for a person.

Plays from YouTube.  [Open on YouTube instead](https://www.youtube.com/watch?v=zED1slvEvU0)

What you no longer do

## Skip the platform work.  Keep the control.

The setup lives in your repository and the builds run on your infrastructure. You see every file Ankra writes and every change it makes.

[Try it on your repository](https://platform.ankra.app/auth/signup)  [See pricing](https://ankra.ai/pricing)

30 vCPU included. Try it without a card; a card unlocks full access and €59 of credit.

- Writing a Dockerfile for Node, Python, Go, Rust, Java, Ruby, PHP, Elixir or a static site
- Paying for CI minutes or running a fleet of build runners
- Putting kubectl or cluster credentials in your CI
- Storing a registry password in your repository
- Installing and wiring a separate GitOps tool

After launch

## Shipping is day one.  Ankra stays for day two.

The same platform keeps your app healthy, reviews your changes and helps when something breaks.

### Pull request previews

Every pull request can get its own preview URL to test before you merge.

### AI review on every pull request

Ankra AI reads each change and posts its findings on the pull request.

### Fix pull requests

When a build breaks, Ankra AI investigates the run and can open the fix for you to review.

### Rollback and promotion

Move a tested image to the next environment, or go back to the last good one.

### Logs, metrics and alerts

See what the app is doing and send alerts to Slack, Teams or your on-call tool.

### Secrets out of your code

Environment secrets live in Ankra or encrypted in Git, never in plain text in your repository.

### See it in the workspace

Three illustrative workflows: launch an application, investigate a failure and prepare another environment.

Ankra  /   northstar

Interactive demo

## You built it.
*Now it’s running.*

A repository becomes an application people can actually use. Ankra takes care of the build, scans and deployment on your infrastructure.

Ship Northstar with Postgres and HTTPS.

1. **Connect the repository**northstar / main
2. **Build and scan the image**Code + container security
3. **Connect the database**PostgreSQL · Stored credentials
4. **Bring the application online**Your cluster · Your domain · HTTPS

The application you built  Now on your infrastructure

northstar.example.com  Live

n  Northstar  A

YOUR WORKSPACE

### Make something
people love.

Ideas are only the beginning.

**The next big thing**   Live and ready to share

**Your next idea**Room to keep building

All systems operational  Built by you.

Application  PostgreSQL  HTTPS

**There’s your app.** Built, scanned and responding over HTTPS.

 HTTP 200

From a repo to a URL you can share.

## Less guesswork.
*More getting on with it.*

Your agent can see what your app is doing, what changed and why it failed. Go from a vague alert to evidence and a fix you can review.

Northstar is failing. Find out why.

1. **Look at the running application**Restarts detected after a deployment
2. **Bring the evidence together**Logs · Events · Deployment history
3. **Find the missing connection**The database secret isn’t referenced
4. **Prepare the configuration fix**A concrete change, ready to review

Ankra investigation  Live application context

northstar / production  Restarting

Application health  ↑ Latest deployment

Container log `Error: DATABASE_URL is not configured`

**Found it. The database connection is missing.**

The credential exists. This deployment doesn’t reference it.

Application environment `− DATABASE_URL: not configured+ DATABASE_URL: secret:postgres-connection`

Ready for your review

**A clear cause. A concrete fix.** Review the proposed change, then choose what happens next.

The context your coding agent was missing.

## Keep the platform.
*Skip the rebuild.*

The stack that works becomes a versioned profile. Reuse it for the next environment, with its own configuration, secrets and cluster.

Give Northstar a staging environment.

1. **Start with your working stack**Application · Postgres · Cache · HTTPS
2. **Save it as a versioned profile**northstar-platform / v1.4
3. **Choose staging’s inputs**Its own cluster, domain and secrets
4. **Prepare the next environment**A stack draft your team can review

One platform, more possibilities  Stack profiles

**Production** Running

Application  PostgreSQL  Cache  HTTPS

Production configuration

northstar-platform **v1.4**

**Staging** Draft

Application  PostgreSQL  Cache  HTTPS

Separate secrets & configuration

Same blueprint **Application + services**

Your new inputs **Cluster · Domain · Secrets**

### Review the staging draft  +

```
profile: northstar-platform@1.4
environment: staging
domain: staging.northstar.example.com
secrets: staging-credentials
```

Reuses the stack definition. Production data stays in production.

**Your next environment starts here.** The platform you already built, with a fresh set of inputs.

Build the platform once. Keep moving.

Your code. Your cloud. Your way forward.  [Make this your workflow](https://platform.ankra.app/auth/signup)

Built for people and agents

## Context. Permissions.
 Execution.

Your coding agent can ship and operate through Ankra too. Ankra connects it to the infrastructure, scopes the tools it can use and records the work it performs.

 [See agents in operation](https://ankra.ai/use-cases/ai-platform-engineer)

### Your infrastructure

An outbound-connected cluster agent links the running system to Ankra.

### Your repository

Stacks, Helm values and manifests keep the intended state in Git.

### Your review

Changes arrive as pull requests and drafts you can read before they apply.

### Your interface

People use the portal, CLI and API. Agents use the platform through scoped tools.

Understand the boundaries

## Your infrastructure. A shared control plane.

Applications and services run on the Kubernetes infrastructure you connect. Ankra’s control plane stores the configuration, credentials and operational context needed to manage it. Read the security and trust pages for the architecture, access controls and data handling.

[Security architecture](https://ankra.ai/security)  [Trust & data handling](https://ankra.ai/trust)

Your next chapter

## You built it.
*Let’s get it live.*

Give your application a path to production.
 Give your team a simpler way to run it.

[Get your app live](https://platform.ankra.app/auth/signup)  [Try a playground](https://ankra.ai/playground)

First 30 vCPU of platform usage included.
Try it without a card; a card unlocks full access and €59 of credit.

From something you built
to something people use.

- Deploy from your repository
- Run on infrastructure you choose
- Operate with your team and agents