Atlas is an open-source database schema management tool by Ariga that treats database schemas as code with a declarative, schema-as-code workflow. Users define their desired schema state in HCL, SQL, or via popular ORMs (Prisma, Django, Hibernate, GORM, TypeORM), and Atlas compares the current database state to the desired state, automatically generating and executing a migration plan to transition between them. Atlas brings the Terraform-style declarative approach to database schema management, making schema changes as reviewable and version-controlled as application code.
Atlas ships with a migration analysis engine that inspects every migration for destructive changes, table locks, potential data loss, and concurrent index violations before deployment. The tool supports two complementary workflows: a declarative mode similar to Terraform where Atlas computes the plan from the desired state, and a versioned mode where migrations are stored as numbered files reviewed in pull requests. Built-in drift detection monitors live databases against the source-of-truth schema and generates exact SQL to reconcile any drift, permission mismatches, or policy violations.
Atlas supports a broad range of databases including PostgreSQL, MySQL, MariaDB, SQLite, SQL Server, ClickHouse, Redshift, Oracle, Snowflake, Spanner, Databricks, CockroachDB, and Aurora DSQL. It integrates natively with CI/CD systems including GitHub Actions, GitLab CI, Bitbucket Pipelines, Azure DevOps, CircleCI Orbs, as well as Kubernetes Operators for Argo CD and FluxCD. Atlas Cloud adds a managed schema registry, visual diff reviews, and policy enforcement for teams running production database migrations at scale.
