One store's problem never affects another
Why every store runs in its own separate environment, and how that separation is checked rather than just promised.
Reqursor runs many stores, for many agencies, on shared servers underneath. Each store runs in its own separate environment, so one store's problem never affects another. That is the guarantee that "shared servers" never means "shared risk": one store's traffic spike, wrong setting, or break-in can't reach another store, or another agency.
What's kept separate
| What's kept separate | What it means |
|---|---|
| Network | Each store gets its own private network space. Nothing reaches it except through the normal path a real visitor would use. |
| Storage | Each store's files and database live in their own storage, walled off from every other store's. |
| Computing power | Each store can only use the share it was assigned: a spike, or a program running out of control, in one store can't leave another without enough. |
| Secrets | Each store's passwords and access keys can be used by that store alone. |
The rule that's never allowed to bend
A store must never be able to reach another store's network, storage, or secrets, and nothing is ever shared that two stores can both change. This isn't an optional setting that someone could turn off by accident. It's built into how every store is made.
How it's enforced, not just promised
- Every check covers it. Keeping stores separate is one of the six areas How Reqursor confirms a store is set up as intended verifies for every store. If a store somehow gained an extra network connection it shouldn't have, that's reported as an unexpected change (a change nobody asked for), not silently allowed to stand.
- Even the AI only sees one store at a time. Every request the AI makes about a store includes only that store's information, never another store's, and never another agency's, even within the same conversation.
- Nothing bypasses the boundary for convenience. Nothing is shared that two stores can both change, not for temporary storage, not for speed, not for anything.
Limits stop one busy store from slowing the others
Every store is assigned a resource profile: a set share of computing power. That limit is enforced, not just a suggestion: one store working unusually hard cannot take away the resources another store was promised, even though they may run on the same machines.
Why this matters for your agency
- Each of your clients stays separate from the others. One client's store having a bad day never becomes another client's problem.
- Damage stays contained. If something does go wrong with one store, the damage stops at that store's boundary: see Why you can always get back to a working store for what happens next.
- A clear answer for questions about compliance. "Every store runs in its own separate environment, and that is verified on every check" is a concrete, checkable claim, not a policy document.
Where to go next
How Reqursor keeps your stores separate
The line between your agency and every other agency (the tenant boundary), which sits on top of this store-by-store separation.
How Reqursor confirms a store is set up as intended
Keeping stores separate is one of the six areas checked on every store, every time.
Core concepts
Definitions of the terms used on this page, like "tenant boundary" (the line between your agency and every other agency) and "secret" (a password or key kept private).
How Reqursor works with your store platform
Why adding support for another store platform does not mean changing the AI. It reads that platform's rulebook instead.
Why you can always get back to a working store
Why recovery is the same safe process as any other change, and how replacing a secret such as a password or key works the same way.