Infrastructure
Provisioning is part of the deploy. Not a separate ops ritual.
Most app teams treat infrastructure as a separate problem from the application — Terraform in one repo, app code in another, an ops engineer holding the keys. Deklarative collapses the two. The same manifest that declares your entities and workflows declares the database they live in, the queue they use, the storage they write to, and the CDN they serve from.
Targets
- AWS — RDS, S3, CloudFront, EKS, Lambda, SQS, SNS
- GCP — Cloud SQL, GCS, Cloud CDN, GKE, Cloud Run, Pub/Sub
- Azure — Azure SQL, Blob, Front Door, AKS, Service Bus
- Hetzner — managed Postgres, object storage, Kubernetes
- Bare-metal Kubernetes — your own cluster, your own rules
- Air-gapped — manifest-driven offline install on Enterprise
Resources
- PostgreSQL clusters with replicas and PITR
- Object storage buckets with lifecycle policies
- Queues — NATS, Kafka, Redis, SQS, Pub/Sub, Service Bus
- CDNs with TLS certificates auto-provisioned
- Container runtimes — Kubernetes, Cloud Run, Lambda, Fargate
- Secrets management — Vault, AWS Secrets Manager, GCP Secret Manager
Reproducibility
- Pulumi state stored in your S3 / GCS / Azure Blob
- Deterministic resource naming with stack + environment
- Drift detection on every apply
- Diff preview before apply
- Rollback to last applied stack
Eject to Pulumi
- Studio generates a normal Pulumi TypeScript program
- You can run
pulumi updirectly without Deklarative - Existing Pulumi stacks importable into the manifest
- No proprietary IaC engine — Pulumi is the runtime
Get started
Describe the system. Ship the system.
Open source under Apache 2.0. Cloud free for evaluation. Production deployments self-hosted or hosted, your call.