DBLift - Database release safety for Python teams
Python-native database release safety
The open core previews SQL, validates history, and undoes a release from the Python workflow you already run. Pro adds SQL review for pull requests. Enterprise adds plan, preflight, and a report your approvers can read.
`$ pip install "dblift[postgresql]"`
Eighteen engines, free to migrate — traditional RDBMS, PostgreSQL-compatible clouds, analytics targets
Oracle
MySQL
PostgreSQL
MariaDB
SQL Server
DB2
SQLite
DuckDB
Azure Cosmos DB
Neon
Supabase
Amazon Aurora PostgreSQL
Google AlloyDB
YugabyteDB
TimescaleDB
Citus
CockroachDB
Amazon Redshift
The release line
From migration intent to reviewable release evidence.
Six steps, one path. Click a step to see what it produces.
plan reads a snapshot model and shows pending migrations, repeatables, checksum drift, affected objects, and validation findings before merge.
Command
`dblift plan --snapshot-model .dblift/prod.json --format html,json,github-actions --output-dir artifacts/`
dblift-plan.html Reviewer report for pending database change.dblift-plan.json Structured plan summary for automation.plan annotations Warnings and blockers surface in the PR.
What review catches
Pick a migration. See what a reviewer would have missed.
Findings land next to the migration line, in the pull request, before the change window opens.
Python workflow
It belongs in the stack you already run.
Install with pip, configure beside the application, run checks in tests or CI. SQL files, Python migrations, reports, and database targets stay on one release-safe path.
`info``migrate``validate``validate-sql`Pro`plan`Ent`preflight`Ent`snapshot`Ent`diff`Pro`export-schema`Pro`undo``baseline``repair``import-flyway`
DBLift Core
Release questions
Each edition removes a different release risk.
OSS Can the team run and reverse migrations from Python projects?
- Same shell and CI job as your application tests
- SQL and Python scripts share one migration history
- Undo included in the open core — CLI plus DBLiftClient
Install now →Pro Can reviewers catch SQL, schema, and data risk before release?
- SQL Review findings for pull requests
- Schema comparison and export for investigation
- Audited Data Fixes with a hash-chained ledger
Buy a license →Ent Can approvers see what will happen and keep the evidence?
- plan and preflight reports from snapshot models
- HTML, JSON, SARIF, and CI annotations
- Guaranteed undo for data corrections when capture is on
pip install "dblift[postgresql]" · Apache 2.0 · 20 engines
Start with the open core. Add review and evidence when risk demands it.
Install DBLift, wire it into your Python project and CI, then compare tiers when reviewers or approvers need stronger controls.