Choose which language model the assistant uses, set how the chat behaves, and see how much it is used, all from a single place.
It sits between the assistant and the language models, so changing a model never means rebuilding anything.
Set up on our own server.
Being built now.
Planned, not urgent.
A thin layer that keeps the assistant independent of any one provider.
In plain language, from a signed-in person in their store.
It uses whichever model and settings are configured.
Requests and usage are tracked for visibility.
In the format the assistant expects, ready to act on.
It grows together with the assistant inside CRM & ERP.
See CRM & ERP