How Qovery Works

Understand the architecture behind Qovery — the Kubernetes control plane for humans and AI agents

The Big Picture

Qovery is the Kubernetes control plane for humans and AI agents — running entirely on your own cloud infrastructure.

Whether a developer deploys via the console, an AI agent deploys via the Agent Skill, or a platform engineer manages clusters via Terraform, they all drive the same control plane. One consistent model, every interface.

How AI Agents Interact with Qovery

AI coding agents (Claude Code, Cursor, OpenCode, and 30+ others) interact with Qovery through two complementary interfaces:

Agent Skill — Forward Engineering

Takes your source code → deploys it on Qovery

The AI agent analyzes your project, generates a Dockerfile, provisions databases, configures environment variables, and deploys — all autonomously. No Kubernetes knowledge required from the developer.

MCP Server — Operations

Manages existing infrastructure via natural language

Once deployed, use the Qovery MCP Server to query environments, troubleshoot deployments, scale services, and manage infrastructure — all through conversation.

Why Kubernetes?

Kubernetes was designed for infrastructure operators — not developers, and certainly not AI agents. Qovery bridges this gap by adding:

• Developer Experience - Deploy without Kubernetes expertise
• AI Agent Interface - Agent Skill and MCP Server for autonomous operations
• Production Readiness - Security, monitoring, and compliance built-in
• Multi-Cloud - Works on AWS, GCP, Azure, Scaleway, on-premise
• Universal Orchestration - Manages containers, databases, Terraform, and more

The Five Products

Qovery’s architecture consists of five integrated products:

Provision

Manage Kubernetes clusters across any cloud. Deploy infrastructure as code with Terraform.

Deploy

GitOps-based deployment for applications, databases, and services. Automatic builds and rollbacks.

Observe

Unified observability with correlated metrics, logs, and events across your stack.

Optimize

Intelligent cost tracking and optimization recommendations for Kubernetes and apps.

Secure

Built-in security and compliance for SOC2, GDPR, HIPAA, HDS, and DORA.

Your Infrastructure, Your Control

Unlike traditional PaaS (Heroku, Platform.sh), Qovery runs on your own cloud accounts:

• Full Ownership - Your infrastructure, your rules
• No Vendor Lock-in - Standard Kubernetes underneath
• Cost Transparency - Direct cloud billing, no markup
• Data Sovereignty - Data stays in your chosen region

Three Layers of Abstraction

1. Multiple Interfaces

Work with Qovery through your preferred method — all driving the same control plane:

• AI Agent Skill - Deploy from Claude Code, Cursor, or any AI coding tool
• MCP Server - Manage infrastructure via natural language
• Web Console - Visual interface for teams
• CLI - Command-line for developers
• Terraform - Infrastructure as Code
• API - Programmatic control

2. Production-Ready Features

Qovery adds what Kubernetes lacks:

• Automatic HTTPS - Free SSL via Let’s Encrypt
• GitOps Workflow - Auto-deploy on Git push
• Preview Environments - Per-pull-request environments
• Database Management - Provision PostgreSQL, MySQL, MongoDB, Redis
• Secrets Management - Encrypted storage with audit logs
• Cost Tracking - Real-time spend visibility

3. Universal Orchestration

Manage any cloud resource from Kubernetes:

Resource Type	Examples
Containers	Docker images, applications
Databases	RDS, Cloud SQL, Azure Database
Storage	S3, Cloud Storage, Azure Blob
Serverless	Lambda, Cloud Functions, Workers
Networking	VPCs, load balancers, DNS
Custom	Any Terraform module

This means one control plane for your entire cloud infrastructure.

Two Workflows

With an AI Agent (Recommended for New Users)

1. Install the Qovery Skill

   Run curl -fsSL https://skill.qovery.com/install.sh | bash in your terminal

2. Open your AI coding tool

   Launch Claude Code, Cursor, OpenCode, or any compatible tool

3. Ask the agent to deploy

   Say “Deploy my application with Qovery” — the agent handles everything from Dockerfile to running deployment

4. Manage with the MCP Server

   Use the Qovery MCP Server to monitor, troubleshoot, and operate your infrastructure via natural language

With the Platform (Console / CLI / Terraform)

1. Connect Git

   Link GitHub, GitLab, or Bitbucket

2. Configure

   Set app settings, variables, and resources via the Console or Terraform

3. Deploy

   Push code → Qovery builds and deploys automatically

4. Scale & Monitor

   Track metrics, scale, optimize — all from one interface

Next Steps

Deploy with AI Agent

From code to deployed in ~10 minutes

All Use Cases

Browse every deployment path

Core Concepts

Learn essential terminology

Platform Products

Deep dive into all five products