Ratify

Upstream schema changes break pipelines. A backend engineer renames a column, their service still works, and nobody thought to check what else depended on it. By the time the data team finds out, the damage is done and the question of who approved what goes unanswered for hours.

Ratify is a data contract workflow engine. When a producer team wants to change a schema covered by an active contract, they raise a proposal in the tool. Consumer teams get notified; what's changing, which columns are affected, whether the change is breaking or additive - and respond before the change ships. The negotiation is tracked. The audit trail is append-only.

Written in Go. CLI-first, using Cobra for the command interface and sqlc to compile explicit PostgreSQL queries into typed Go. React and TypeScript for the web UI. Ships as a single binary with no runtime dependencies, a Docker image on Docker Hub, and via GitHub Releases.

Proposal workflow

Key decisions

Decision	Chosen	Why
Backend language	Compiled Go	Single static binary, zero runtime dependencies, goroutines for concurrent schema introspection
Database access	Plain SQL + sqlc	Bypasses ORM abstractions and N+1 overhead — compiles explicit PostgreSQL queries into type-safe Go
HTTP router	Chi	Strict compatibility with net/http standard library, lightweight, clean middleware composition
Metadata store	PostgreSQL	Raw sequentially numbered migrations via golang-migrate executed natively at startup
Task scheduling	Embedded robfig/cron	Eliminates distributed worker dependencies (Temporal, Celery), in-process cron running hourly schema checks
Auth	API Keys + JWT	API keys for machine access, short-lived JWTs for web sessions. No OAuth callback complexity.
Consumer responses	Token-hashed links	Account-less interaction via cryptographically secure random tokens, consumers respond without creating an account
License	MIT	The OSS core needs to be trusted before any monetization can happen