FOR MANAGED POSTGRESQL · NO SUPERUSER · NO WAL ACCESS

A backup you haven’t restored is a rumor.

PostGrab takes scheduled logical backups of your managed Postgres — DigitalOcean, Supabase, Neon, RDS, anything pg_dump can reach — then proves each one by restoring it and checking what came back. You get the receipt, not a promise.

Free tier: one database, daily backups, verify any of them with a click. No card.

VERIFICATION RECORDVERIFIED ✓
targetapp-prod · postgres 18
started2026-10-26 03:00:04 UTC
dumppg_dump · 4.2 GiB compressed
sha2569f3c1b…e2a71d ✓ recorded
restoreinto scratch · 212 tables
rows1,284,112 restored
verifymatches source ✓
elapsed6 m 41 s
next run · tonight 03:00 UTC

Works with DigitalOcean · Supabase · Neon · Amazon RDS · Heroku Postgres · any Postgres you can reach over TLS

THE GAP

Managed Postgres left your backups to someone else.

The serious tools — pgBackRest, WAL-G, Barman — need the machine: filesystem access, WAL archiving, superuser. Your provider gives you none of that. What’s left is their snapshot button.

Locked to the provider

Snapshots restore whole clusters, inside the same provider, on their terms. You can’t pull one to your laptop, your CI, or anywhere else.

Never actually tested

A snapshot is a promise that a restore would probably work. When did anyone on your team last watch one succeed?

Retention isn’t yours

Seven days, thirty if you’re lucky. Your obligations and your provider’s defaults are two different documents.

PostGrab needs exactly one thing: a connection string.

HOW IT WORKS

Three steps, then receipts forever.

01

Point it at a database

Paste host, database, user. The password is sealed to the backup workers’ public key the moment it arrives — the control plane stores a blob it cannot open.

02

Pick a cadence

Nightly at 03:00, every six hours, hourly. Schedules live server-side and fire whether your laptop is open or not — and you’re told when one is missed.

03

Get proof, not logs

Every run is dumped with pg_dump, checksummed and stored. Verify any backup with an actual restore, in one click. The dashboard shows what was proven, and when.

RESTORES

The restore button is the product.

Everything else exists so that this works on the day you need it.

Download any backup

A plain pg_dump archive. Yours, portable to any Postgres, anywhere.

Verify on demand

One click restores a backup into a scratch database and checks what came back. No production risk.

Restore in place

Push a chosen backup into a target you control — deliberate, audited, never automatic.

Verified means verified

A backup is only marked ✓ after a real restore succeeded — not after the dump exited zero.

BACKUPS · APP-PROD

when (UTC)sizestatus
10-26 03:004.2 GiBverified ✓
10-25 03:004.2 GiBstored
10-24 03:004.1 GiBstored
10-23 03:004.1 GiBverified ✓

retention · 7 daily · prune is a dry run unless applied

SCHEDULED RESTORES

Staging rebuilt from last night’s production. Every morning.

A scheduled restore takes the latest successful backup of one database and restores it into another you nominate — replacing its contents each time. Point it at staging, and every workday starts with yesterday’s real data instead of a hand-rolled script somebody resents maintaining.

It also quietly answers the only question that matters: if staging came up this morning, your backups restore. You stop taking anyone’s word for it — including ours.

SOURCEapp-prod
backup · 03:00
POSTGRABlast night’s backup ✓
restore · 05:00
DESTINATION · REPLACEDapp-staging

Destinations are explicit and never your source — PostGrab refuses a schedule that would restore a database onto itself.

SECURITY

Built as if it holds production credentials. Because it does.

The honest version, not the badge wall:

Sealed credentials

Database passwords are encrypted to the backup workers’ key on arrival. The API that stores them cannot decrypt them — and neither can anyone who steals its database.

Postgres-enforced isolation

Tenants are separated by row-level security in Postgres itself — not by WHERE clauses we promise to remember.

Append-only audit log

Every change — databases, schedules, members, sign-ins — lands in a log that nothing in the product can edit or delete.

Signed webhooks

Notifications are HMAC-signed, so your receiving end can prove who’s calling before it trusts a word.

Two-factor sign-in

TOTP with recovery codes. Sessions are opaque tokens, and the database stores only their hashes.

Honest about scope

Logical backups mean granular and portable — not point-in-time recovery. We document exactly what you get, and what you don’t.

PRICING

Start free. Pay when it matters.

Free

$0 /mo

  • 1 database
  • 5 GiB stored
  • Daily backups
  • Verify any backup with one click
  • Restores and downloads included
Create an account
MOST POPULAR

Standard

Pricing at launch

  • 5 databases
  • 100 GiB stored
  • Backups every 6 hours
  • Scheduled restores
Create an account

Pro

Pricing at launch

  • 25 databases
  • 1 TiB stored
  • Hourly backups
  • Everything in Standard
  • Priority support
Create an account

Every tier can verify any backup, and every tier has unlimited team seats with roles — proof and a second admin aren’t paid features. Paid plans are arranged in-app after you sign up.

Stop hoping. Start checking.

Create an account, point PostGrab at a database, and watch the first backup land tonight. Then verify it with a click.

Create an account

free tier · no card · one connection string