Ship SQL to production the way you ship code
Sql2DevOps runs your SQL as versioned packages from development to production, with approvals, a tracked emergency path and a tamper-evident history your auditor can verify.
SQL Server · PostgreSQL · SQLite · Oracle and MySQL coming soon · on-premises or in the cloud · GitHub Actions, GitLab CI, Azure DevOps, Jenkins · one binary for Windows, Linux and macOS
$ sql2devops promote CHG-1001 --project finance CHG-1001 v1 (sha256:9f2c41e07b6a...) to PROD/finance-db step 1/3: 01-rates-table.sql (ddl) ............ ok 41 ms step 2/3: 02-interest-function.sql (other) .... ok 12 ms step 3/3: 03-initial-rates.sql (data) ......... ok 8 ms succeeded · recorded as event #12 · signed by ci-runner (a1c94e0b72d3) # approved by github:maria, after the same version succeeded in DEV and HML
Sound familiar?
Application code has pipelines, reviews and history. The database usually gets a script pasted into a console at 11 p.m.
"Did this run in production?"
Nobody knows which scripts reached which environment, in which order, or who ran them.
"It worked in staging"
Production received a slightly different script than the one that was tested.
"The fix was urgent"
Emergency changes skip every rule and are forgotten, until the next deployment breaks.
How it works
Three ideas: packages, promotion and a history nobody can quietly rewrite.
Package your change
A folder with your scripts, their order and their rollbacks. Submitting it stores an immutable copy with its hash.
package: CHG-1001
target: finance-db
steps:
- file: 01-rates-table.sql
rollback: rollback/01.sql
Promote it
The same version moves DEV → HML → PROD. Each environment has its own rule: no approval, the pipeline's evidence, or named people.
$ sql2devops promote CHG-1001 $ sql2devops approve CHG-1001 \ --env PROD
Keep the proof
Every submission, approval and run is a signed, hash-chained event in your own Git repository. One command checks it was not altered.
$ sql2devops verify consistent: 248 events, 2 projects, all signed
See everything at a glance
The web interface (paid edition) shows every package across your environments, what is waiting and why, each run with its steps and the exact line that failed.
| Package | Version | DEV | HML | PROD | Next |
|---|---|---|---|---|---|
| CHG-1001 | v1 | ok | ok | ok | done |
| CHG-1002 | v2 | ok | ok | approved | PROD (ready) |
| CHG-1003 | v1 | ok | failed | – | HML: fix and re-run |
| HOTFIX-7 | v1 | – | – | ok* | review; deploy to DEV, HML |
Built for the day something goes wrong
Your scripts, your rules. Flexible by default, strict where you configure it, and everything recorded.
Failure reports that help
The script, batch and exact line, the server's message and the surrounding code, and the state the database was left in.
Rollbacks per step
On failure, the failed step and the completed ones are rolled back, newest first. Or stop and keep everything for analysis.
Emergencies, tracked
Bypass the rules when production is down, with a reason and a ticket. Follow-ups stay visible until reviewed and reconciled.
Your Git, your data
History lives in a Git repository you own. No server, no account, nothing sent to us.
Secrets stay secret
Passwords enter scripts as variables from your pipeline's secrets and are masked in every report.
One small binary
No JVM, no runtime. Download one file, or use the ready-made GitHub Action.
Your databases, on-premises or in the cloud
Sql2DevOps talks to each engine with its own native protocol, so the managed versions in the cloud work exactly like your own servers.
Available
- Microsoft SQL Server
- PostgreSQL
- SQLite
Coming next in development
- Oracle Database
- MySQL
- MariaDB
In the cloud
- Azure SQL Database and Managed Instance
- Azure Database for PostgreSQL
- Amazon RDS and Aurora (SQL Server, PostgreSQL)
- Google Cloud SQL and AlloyDB
- Supabase, Neon and other PostgreSQL services
- Oracle Autonomous Database, RDS for Oracle and MySQL: with the new engines
Using another database? Tell us which one: it helps us decide what comes next.
Simple pricing
Start free. When your team needs the whole thing, one flat price per company, not per user, not per database.
Free
- Executor for SQL Server, PostgreSQL, SQLite
- Signed, verifiable history in your Git
- Approvals and tracked emergencies
- Packages with up to 2 scripts
- Projects with up to 2 environments
- Web interface
- Email support
Pro
- Everything in Free
- Unlimited scripts per package
- Unlimited environments
- Web interface
- Email support
- License checked offline, nothing sent anywhere
Larger organizations, invoices or a yearly contract? Talk to us.
Get started in a minute
Get your free download key on the download page, then install with one command and run your first package.
# Windows (PowerShell) PS> $env:SQL2DEVOPS_DOWNLOAD_KEY="s2dk_..."; irm https://sql2devops.com/install.ps1 | iex # Linux / macOS $ curl -fsSL https://sql2devops.com/install.sh | SQL2DEVOPS_DOWNLOAD_KEY=s2dk_... sh # Pipelines (GitHub Actions shown; works in any CI) - run: curl -fsSL https://sql2devops.com/install.sh | sh env: SQL2DEVOPS_DOWNLOAD_KEY: ${{ secrets.SQL2DEVOPS_DOWNLOAD_KEY }} SQL2DEVOPS_VIA: action
Questions
Does Sql2DevOps check or rewrite my SQL?
No. It runs your scripts exactly as written, in the order you define, and reports precisely what happened. What the scripts do, and their rollbacks, is up to you.
Where is my data?
Nowhere but your own infrastructure. The history lives in a Git repository you own, and the tool never contacts our servers, not even to check the license.
Can I adopt it on an existing database?
Yes. Package your next change and start from there; scripts that already ran by hand can be recorded so later deployments know about them.
What happens when the license expires?
The free edition applies again. Limits are checked before anything runs, so a deployment is never interrupted halfway.
Is it open source?
The source is available under the Business Source License 1.1: you may use it in production, and each version becomes Apache 2.0 four years after release.