One schema change.
A clear path to code review.
Connect your SQL database to GitHub. AI-assisted workflows turn schema changes into reviewable code changes.
Connect your database
Start where your schema lives. Connect PostgreSQL, MySQL, or Microsoft SQL Server.
Learn what data we use and how we protect it in our Security & data guide.
ALTER TABLE customers
ADD COLUMN display_name TEXT;A change to the public schemaFollow your repository branch
Link a branch and add project instructions. The workflow uses your schema changes to propose code updates.
Use SQLAlchemy 2.0 typed models.
Preserve API field names, make new nullable columns optional, and add tests for changed models.
Do not edit existing migrations.
Keep the review in GitHub
Inspect the schema diff and the proposed code. Your team decides what gets merged.
+ display_name: str | NoneProposed model change, ready for reviewIllustrative schema, repository, and code examples. Actual proposals depend on your schema and project.
Find opportunities in the schema you already have.
Schema optimisation reviews your recorded database schema alongside your codebase, looking for meaningful indexing and data-model improvements.
Follow optimisation runs in Run history, then review proposed changes and their reasoning in a GitHub pull request.
Connect your databaseCREATE INDEX orders_customer_id_idx
ON orders (customer_id);If the application frequently retrieves orders by customer, an index on customer_id may reduce full-table scans. Review query patterns and write overhead before applying it.
A shared workflow.
A clear owner.
Keep databases, repository access, and run history together in your organisation. Invite teammates with owner, admin, and member roles.
Set up your organisation- Organisation-owned resources
- Switch workspaces without mixing up your team’s databases and repositories.
- Visible progress and outcomes
- Follow runs, inspect schema diffs, and open the resulting GitHub pull requests.
- Configuration in context
- Choose schemas and give each branch connection the instructions it needs.
Start at the source.
Connect a database. Choose a branch. Review the changes.
See pricing