Which memory this chat reads, including none.
Off is a real option with its own card and its own sentence - every AI starts from nothing, and nothing new is saved. A memory control without an off switch is a setting, not a choice.
The explanation is the product's actual argument: most apps keep what they learn inside one company's model, and this one keeps it separately so every AI reads the same notes. That is why switching model mid-chat costs nothing, and it is worth stating rather than leaving as a feature name.
import { MemoryScope } from '@redrob-labs/ui';
| Prop | Type | Notes |
|---|---|---|
options |
MemoryScopeOption[] |
|
value |
string |
|
defaultValue |
string |
|
label |
string |
|
onTitle |
React.ReactNode |
|
offTitle |
React.ReactNode |
|
offText |
React.ReactNode |
|
summary |
React.ReactNode |
|
lede |
React.ReactNode | false |
|
foot |
React.ReactNode |
|
compact |
boolean |
Tightens the panel for a composer popover: a smaller status card and less space between parts. |
slotClassName |
{ card?: string; lede?: string; options?: string; foot?: string } |
Class names added to the panel's parts, so a consumer restyles one without overriding the whole panel. |
className |
string |
|
onChange |
(value: string, option: MemoryScopeOption) => void |