Test mode first: earning trust before writing anything

Software that writes to your store should have to earn that right. Reloop is built around a simple idea: you should be able to watch it think for weeks before you let it act.

Rehearsals, not simulations

In test mode, Reloop does the entire job except the final write. It imports your subscriptions, counts every delivery, plans every gift, and then rehearses each one against live data: is the subscription active, is the renewal date right, does the gift product still exist, would the write succeed.

Every rehearsal produces a preview of the exact gift line it would create, down to the field. You can open any queued gift and read precisely what would happen and when. Nothing is written anywhere.

Going live changes one thing

When you switch to live, the checks do not relax. The same rehearsal runs immediately before every write, only gifts whose product links have been verified are eligible, and every write is read back and confirmed field by field before it counts.

Anything that cannot pass its checks simply stays queued with a plain explanation of what is blocking it. Anything unexpected stops and appears in your attention list. The system prefers asking to guessing, every time.

Why we built it this way

Reloop runs against real stores with real customers, where a wrong write is not a bug report but somebody's order. Building the cautious version first meant our first ever live gifts went out verified, on the right renewals, at a price of zero, with an audit trail for every decision. That is the only way we would want software behaving inside our own store, so it is how Reloop behaves in yours.