Deploy with AI Agent

Tell your AI coding agent to deploy. It handles everything — from your code to a live Kubernetes deployment.

Deploy to Kubernetes with your AI agent.

Qovery account — console.qovery.com (free to start)

Git repository — GitHub, GitLab, or Bitbucket

One install. Your agent picks the right skill automatically.

curl -fsSL https://skill.qovery.com/install.sh | bash

→ All skills

Then open Claude Code, Cursor, OpenCode, or any compatible tool and say:

claude

Deploy my application with Qovery

No Dockerfile. No Kubernetes knowledge. No cluster needed upfront.

---

Compatible Tools

Claude Code

Anthropic · Terminal Recommended

OpenCode

OpenCode · Terminal Open source

Gemini CLI

Google · Terminal

Cursor

Anysphere · IDE

VS Code Copilot

GitHub · IDE Extension

Roo Code

Roo · IDE Extension

Kiro

Amazon · IDE

Junie

JetBrains · IDE Extension

Goose

Block · Standalone Open source

Amp

Sourcegraph · Standalone

OpenHands

All Hands AI · Standalone Open source

OpenAI Codex

OpenAI · Standalone

Mistral Vibe

Mistral · Standalone

TRAE

ByteDance · Standalone

Any tool discovering skills from .claude/skills/, .config/opencode/skills/, or .agents/skills/ is compatible.

8 skills, one selector — pick your use case

qovery-deploy · qovery-preview · qovery-optimize · qovery-troubleshoot · and 4 more →

How Deployment Works

When you ask your agent to deploy, qovery-deploy runs through this workflow autonomously:

1. Analyzes your codebase

   Detects language, framework, ports, database needs, and environment variables.

2. Creates a Dockerfile if missing

   Production-ready multi-stage template for your stack. Custom-generated for any other framework.

3. Asks the right questions

   Dev or production? Which database? CLI+API or Terraform? A few questions — no YAML to write.

4. Provisions infrastructure

   Cluster if needed, databases in container or managed mode, env vars with aliases and overrides.

5. Deploys and watches

   Monitors the rollout. Auto-fixes build errors, port mismatches, health check failures. Asks before touching your code.

---

How It Compares

	Console	AI Agent Skill	MCP Server
Creates Dockerfiles	— No	✓ Yes, 12+ frameworks	— No
Provisions databases	Manual	✓ Automatic	— No
Monitors & fixes failures	Manual	✓ Auto-fix	Query-based
Best for	Fine-tuning	Deploying from code	Day-2 ops

The Skill deploys. The MCP Server operates. Use both together. → MCP Server

---

Project-only install: curl -fsSL https://skill.qovery.com/install.sh | bash -s -- --project

Update: re-run the install command.

---

Next Steps

All Skills

Deploy, troubleshoot, optimize, preview

MCP Server

Manage infrastructure from your AI agent

Terraform Provider

Infrastructure as Code

Skill Source Code

Fork or contribute