Personalization & Memory

Teach the assistant who you are — profile details and saved memories that personalise every response.

Personalization (My Settings → Personalization) stores what the assistant should know about you, so answers arrive pre-personalised instead of you repeating context in every chat.

Profile

FieldPurpose
Full nameHow the assistant identifies you.
What should IntraLLM call you?Preferred name/honorific — "Dr Patel", "boss", "sam".
Save ProfileApplies your changes.

Profile context shapes tone and formality across chats.

Memory

Memory holds discrete facts the assistant recalls in conversation — your role, standing preferences, recurring project context.

Add a memory item

Type a short factual statement, for example:

I manage the Sydney data-centre refresh project.
I always want temperatures in Celsius and dates DD/MM/YYYY.

Select Add a memory item. It appears under Saved memories with a running count.

Manage memories

ActionEffect
DeleteRemove one memory.
Clear AllRemove every stored memory.
Rebuild IndexRe-index memories so recall stays accurate after many changes.
Enable memory in chatMaster switch — off means no memory is used.

What memory is good (and not good) for

Good usePoor use
Stable facts: your role, region, format preferencesFast-changing data (tomorrow's schedule)
Style preferences: bullet points, spelling localeSensitive personal data you shouldn't store — it persists
Recurring project context for many chatsCompany-wide knowledge — that belongs in a knowledge base

Note:

Memories are per-user and stored in your account. Keep them factual and current — stale memories can mislead answers as easily as fresh ones help.