Webinar Sept 24: Heroku to AWS in one command, with an agent doing the work.

Deploy goes
beyond ArgoCD.

Containers, Terraform modules, managed databases, Helm charts - deployed together, rolled back together, previewed per PR. The unit of deploy is the environment, not the pod.

Get startedTalk with us
1 graph
apps + infra
5+ runners
CI/CD agnostic
< 30s
preview deploys
1-click
rollbacks
Why this layer exists

Apps and infra live in two worlds.

Your app deploys from ArgoCD. Your database from Terraform. Your third-party service from a custom script. Rollback means reverting all three, in the right order, at 2am. Deploy is the layer that treats the whole environment as one unit.

01

Apps + infra are two worlds

Terraform for infra, ArgoCD for apps, a script for the rest. Rollback means reverting all three in the right order.

02

Every team picks a different CI

GitHub Actions, GitLab CI, Jenkins, CircleCI - your platform team writes adapters for all of them. Forever.

03

External services are first-class

Redis, Postgres, S3, a third-party webhook - all part of your service graph. None of them deploy the same way.

What Deploy gives you

The whole unit,
shipped.

Six capabilities that unify everything your team ships into one deployable graph. Apps, databases, Terraform, Helm - previewed, deployed, rolled back together.

01Apps, databases & jobs
Containers, managed databases, cron jobs, lifecycle jobs - all in the same environment definition. Deployed together.
02Terraform, inline
Terraform modules are first-class. Deploy them alongside your apps. State managed. No separate pipeline.
03External services
Redis, S3, Stripe webhooks, any managed resource - declared, provisioned, and wired into your service graph.
04CI/CD agnostic
GitHub Actions, GitLab CI, Jenkins, CircleCI - or no CI at all. Qovery builds and pushes. Your CI can still trigger.
05Ephemeral previews
Every PR gets its own full-stack environment. Apps, databases, seeded data. Merged → gone. < 30 seconds typical.
06Canary, blue-green, rollback
Progressive delivery strategies out of the box. One-click rollback that reverts the whole environment, not just one pod.
Works with

Your pipeline,
still yours.

Deploy does not replace your CI. It replaces the part that nobody wants to maintain - the deployment pipeline that glues build to cluster.

GitHub ActionsNATIVE
Trigger on push, PR, or tag. Qovery deploys after build.
GitLab CINATIVE
Same pipeline, Qovery as the deploy step. No rewrites.
ArgoCDBRIDGE
Existing ArgoCD? Qovery can orchestrate alongside, not replace.
Jenkins / CircleCI
Webhook trigger. Call the API. Qovery handles the rollout.
Terraform
Inline modules or external state. First-class, not bolted on.
Helm
Helm charts deployed as services. Values per environment.
Docker / OCI
Any OCI image. Buildpacks or Dockerfile. Your registry or ours.
Nx / Bazel
Monorepo support. Affected-only deploys. Change detection baked in.
AI Agentic Layer

The agent
ships the unit.

The Agentic layer treats the environment as one deployable thing. Ask it to add a cache, wire a webhook, clone prod - it drafts the plan, you approve.

$ billing service needs Redis cache + Stripe webhook endpoint. Add to staging, test, then roll to prod.
✓Add Redis 7.2 to staging (elasticache.t4g.micro · $12/mo)
✓Wire STRIPE_WEBHOOK_SECRET from Vault → env var on billing-svc
✓Register endpoint https://staging.acme.qovery.app/webhooks/stripe
✓Run integration tests · 14/14 passed · Stripe test events received
✓Promote staging → prod · awaiting approval · rollback plan attached
FIX

Fixes the graph, not the pod

The agent understands that adding Redis means updating the app config, the network policy, and the cost forecast. One prompt, all three.
WIRE

Wires dependencies for you

Secrets, environment variables, service discovery - the agent writes the glue that your team hates writing manually.
SAFE

Tests before prod

Deploy to staging, run the tests, show the results, then ask for prod approval. The same workflow your team already trusts, now automated.
"Our deployment times went from 55 minutes to 8 minutes. It's a night-and-day difference."
Jean-Baptiste Barth, Infrastructure Lead · Alan
The other layers

Provision

Kubernetes management

Observe

Monitoring & observability

Optimize

Cost & performance

Secure

Governance & compliance

Ship the whole unit,
not the pods.

Connect your repo and your cluster. Deploy apps, databases, Terraform modules, and Helm charts as one environment. Preview per PR. Rollback with one click.

Deploy your first appRead the Deploy docs