The best Supabase backup tools in 2026

Six ways to back up a Supabase project, from the built-in backups to dedicated services to a cron job you write yourself. One disclosure up front: BackupDrill is our product — it is in the list, clearly marked, and judged on the same dimensions as everyone else. Every fact below was checked against each tool’s current pricing page and docs in July 2026.

At a glance

Facts and pricing as of July 2026 — check each tool’s site for current numbers. One scope note that applies across the board: Supabase restricts access to its managed schemas, so database backups here — Databasus’s and BackupDrill’s included — cover the public schema by default, not auth or storage internals.

Supabase built-in backups and PITR

The baseline everything else gets measured against. On the Pro plan ($25/month) and above, Supabase takes a daily backup of your database automatically — 7-day retention on Pro, 14-day on Team — and daily backups can be downloaded while the project stays on logical backups (projects on Postgres 15.8.1.079 or newer, plus anything with PITR, use physical backups, which can’t be downloaded). When a day of possible data loss is too much, the official upgrade is the point-in-time recovery add-on: from $100/month per project for a 7-day window, recovery to any second, worst-case RPO around two minutes. It requires at least a Small compute add-on and replaces the daily backups once enabled.

The limits, from Supabase’s own docs: backups do not include Storage files (only their metadata), backups live inside the platform with retention tied to your plan, PITR offers no export, and nothing ever restore-tests what was captured. We compare the PITR trade-offs in detail in Supabase PITR vs BackupDrill.

Choose it if: you are on Pro anyway and want zero setup — you already have this. Add PITR when losing up to a day of writes is unacceptable; it is genuinely the best recovery granularity on this page.

Databasus

The strongest open-source option: Apache 2.0, free, self-hosted via Docker, around 7.6k GitHub stars as of July 2026. It backs up Postgres 14–18 (plus MySQL, MariaDB, and MongoDB), connects to Supabase over the Session Pooler, and — credit where due — it restore-tests its own backups: it periodically restores the latest backup into a throwaway database container and reports the outcome. It also does WAL streaming for point-in-time recovery and physical and incremental backups (PostgreSQL 17+), which nothing else in its price class offers — though only for databases it can reach directly: Supabase-managed projects connect over the Session Pooler and get logical dumps, no PITR.

The limits: it is database-only — no Supabase Storage file backup, so uploads need a separate answer — and self-hosted means you run it: the server it lives on, its updates, and the monitoring that tells you the backup machine itself is still alive. Per its own docs, the Supabase backup covers the public schema. The full head-to-head is in Databasus vs BackupDrill.

Choose it if: you are comfortable operating your own infrastructure and want free backups with real restore verification — and your Storage files either do not matter or are covered by something else.

SimpleBackups

The broadest hosted coverage of the Supabase stack. Its OAuth integration backs up the database, Storage files, Edge Functions source, Auth configuration, PostgREST config, and secret names in one connected setup — more of the platform surface than anything else on this page, BackupDrill included. Backups can go to any storage provider you own, or theirs. The free tier covers one backup job; paid plans start at $49/month (Lite: 5 backup jobs, 12-hour frequency) and run to $299/month, with yearly billing about 20% cheaper.

The gap: no automated restore verification on any tier. Restoring means following their documented commands yourself, and the disaster recovery dashboard on higher tiers is audit and visibility, not restore testing. The full head-to-head is in SimpleBackups vs BackupDrill.

Choose it if: you want the widest possible slice of your Supabase project captured — Edge Functions and Auth config included — or you have many backup sources beyond Supabase and want them in one dashboard.

supabackup

The minimalist of the group: it dumps your Supabase database to your own Google Drive on a schedule, and that is the whole product. Free plan: one weekly backup job. Pro: $9/month for three jobs, daily backups, and retention control (keep the latest 7, 14, or 30 backups, or keep everything). Setup is signing in, connecting Google Drive, and pasting a database connection URI.

The limits are the flip side of the simplicity: database only — no Storage files — no restore verification, and Google Drive is the only destination. As a cheap “at least the data exists somewhere else” layer for a side project, that can be exactly enough.

Choose it if: you have a side project, want the cheapest hosted daily dump, and a folder in the Google Drive you already pay for sounds better than configuring an S3 bucket.

GitHub Actions (DIY)

Supabase officially documents this path: a scheduled workflow running supabase db dump that commits SQL files to a private repository — their docs warn twice, never a public one. The common variant is a cron workflow running pg_dump and pushing the archive to S3 or R2; our step-by-step GitHub Actions guide has a complete workflow, and the backup options guide has working commands. Cost: effectively free — runner minutes plus Supabase egress.

What you own in exchange: keeping pg_dump version-matched to your project’s Postgres major, secrets handling, alerting when the job fails — the classic failure mode is a cron that has been silently red for a month — plus scripting Storage file backups yourself, and scripting restore testing yourself if you want to know the dumps actually restore.

Choose it if: engineering time is cheaper for you than a subscription and you will genuinely maintain it — including the unglamorous parts like failure alerts and version pinning.

BackupDrill — disclosure: this is our product

We built BackupDrill around one question the rest of this list mostly leaves open: after the backup runs, does anyone prove it restores? It backs up the database (public schema by default, like the other pg_dump-based tools here) and Storage files together into your own S3, R2, or B2 bucket — that is the only mode; there is no bundled storage — with a manifest of checksums alongside. Then, on a schedule, it runs a restore drill: the latest snapshot is restored into a throwaway Postgres and verified against the manifest, including sha256 checks on a sample of the backed-up Storage files, and the graded report lands in your inbox.

Pricing: Free backs up one project weekly with one drill on the first backup. Solo is $19/month for five projects, daily backups, and monthly drills; Team is $49/month with weekly drills; Agency is $99/month and adds client-ready PDF reports. The backup engine is an open-source MIT CLI (github.com/backupdrill/cli), so restores work with stock pg_restore with or without us. What it does not do, so you do not have to find out later: no point-in-time recovery (snapshots only), no Edge Functions or Auth config backup, Supabase-hosted projects only.

Choose it if: you want the database and its files backed up to a bucket you control, and you want standing proof the backups restore — without being the person who has to remember to test them.

FAQ