Drop your prod table, click “Replay”. A one-line tweak makes it possible.
Has your LLM deleted your production db? Do you still cross your fingers every time you run ALTER TABLE in prod? Do schema tweaks force you into late-night migrations and manual patch jobs? Wish you could redesign a table, truncate it, and restore the data with one click?
The mental shift is that you emit an event BEFORE you write to your database. That way the event becomes the truth of what happened. Flowcore automatically stores the event, in the order it came in, and immediately sends the event back to your application which is when you write it to your database.
So you as an application developer only need to create a completely normal POST endpoint, Flowcore automatically sends the event to that endpoint.
This separation, storing an event before you store an entry in the database, makes the file where you write to your database into something that we like to call a transformer because it transforms immutable events into the concrete rows your application actually queries.
When you hit Replay, Flowcore re-streams every past event to that POST endpoint, and the transformer runs through them in order. Re-inserting, updating, or deleting rows exactly as it did the first time so your tables return to a perfectly consistent state without a single migration script or manually written back-fill job. It’s like you’ve stored every CRUD operation with it’s original payload of data and now it’s being reapplied in the same or a new way.
