Sort

Sort's core idea comes straight from the pull-request habit. A user proposes a change, an admin or engineer approves or denies it, and approved changes get applied to the database for you. It groups several changes to multiple tables into a single request, so a reviewer can clear the whole set in one pass rather than clicking through row by row.

Once a request is approved, Sort can write the change to the underlying Postgres or Snowflake database itself, so you don't need SQL expertise to land it. The change runs against the real schema, which keeps hand-written scripts and copy-paste mistakes out of the loop.

Sort logs every request, review, and execution in one place. That history answers the questions teams usually can't: what changed, who approved it, and why. For anyone who has ever hunted through database logs after a bad edit, the paper trail is the point.

Alongside changes, Sort carries issue tracking. A teammate can flag a data problem, assign it to the right person, and follow it to resolution, which keeps known data gaps from getting buried under day-to-day work.

The Sort API lets teams file and track issues and change requests from their own tools, not just the web UI. You can wire it into scripts or internal apps so an automated process proposes an edit and a person still approves it before it applies.

Sort includes data discovery plus a Markdown interface for documentation. Teams can write docs that hold example queries and run them from the same window, so the catalog stays readable and connected to the live database instead of drifting into a separate wiki.

Sort separates who can propose changes from who can approve them, and admins set those roles. That split is what lets a frontline team move fast on fixes while keeping a small number of people accountable for what reaches production.