Adjoint

Adjoint gives every developer their own production-like Postgres database in seconds — a full-size, copy-on-write clone of a curated golden snapshot, running on AWS infrastructure your organization owns.

No shared staging database. No multi-hour restores. No stale sample data. Each developer clones, breaks, resets, and destroys freely, in a namespace that's theirs alone.

How it fits together

Adjoint is two planes with a hard boundary between them:

• Control plane — Adjoint-hosted. Stores orgs, users, the golden-snapshot catalog, and audit history. It orchestrates, but it never touches or stores your data.
• Data plane — your AWS account. An EKS cluster backed by FSx for OpenZFS, running CloudNativePG (CNPG) Postgres instances. This is where copy-on-write cloning actually happens, and where your data lives — under your AWS credentials, in your VPC.

An admin curates versioned golden snapshots from a real source of truth (e.g. RDS Postgres). Snapshot versions are immutable. Every developer clone is a ZFS copy-on-write branch of one of those versions: full data, isolated changes, near-zero marginal storage cost.

How it feels

$ adj join adj1.eyJ2IjoxLCJzIjoiaHR0cHM6...   # the one-line invite from your admin
$ adj clone prod-pg16 --wait
Instance 0ab2b1ab-bd4a-4e1a-9ab8-c3b0251ee37d is Ready at i-0ab2b1ab...-rw.ns-11d3....svc:5432
$ psql "$(adj creds 0ab2b1ab-bd4a-4e1a-9ab8-c3b0251ee37d --uri)"
psql (16.4)
Type "help" for help.

app=>

That's the whole loop: join the org once, then clone, connect, and work.

Two tracks

• Admin: provision + curate — bootstrap the AWS data plane, connect a source, import a golden snapshot, invite your team.
• Developer: clone + work — join an org and get a working database in about two minutes — or seconds, if your admin runs a warm clone pool.

Not installed yet? Start with Install.