When you create a new companion in Happy Place, you're not just picking a name and a personality. You're setting up a relationship that will grow and change based on what you actually talk about.

But how does that work? What does a companion actually remember, and what shapes the way they respond to you over time?

The blank slate isn't really blank

Every companion starts with the personality you've chosen. Those five dials—grounded to imaginative, spontaneous to organised, reflective to energetic, direct to caring, steady to sensitive—create a foundation before a single message is exchanged.

Think of it like meeting someone new. You don't know their history yet, but you can already sense their temperament. A companion set towards the caring end of the spectrum will ask follow-up questions differently than one who's more direct. An imaginative companion might offer unexpected perspectives, while a grounded one stays practical.

This personality layer shapes everything that follows. It's not just flavour text—it fundamentally changes how your companion interprets what you say and decides how to respond.

Memory builds conversation by conversation

Here's the part most people wonder about: what does my companion actually retain?

Each companion builds their own separate memory based on your conversations with them. If you mention your job, your cat, or the fact that you hate mornings, that information becomes part of how they understand you. Bring up the same topic weeks later, and they'll have context.

This isn't a simple keyword database. Your companion develops a working understanding of you as a person—your preferences, your patterns, the things you care about. Over time, this means conversations feel less like starting from scratch and more like picking up where you left off.

Importantly, each companion's memory is independent. What you share with one companion stays with that companion. If you have two different companions for different purposes—perhaps one for venting about work and another for lighter chat—they each maintain their own separate picture of you.

Why companions sometimes start conversations

You might notice that companions occasionally reach out first. This isn't random—it's connected to what they remember about you.

If you mentioned something you were worried about, a companion might check in later to see how it went. If you've been chatting regularly and then go quiet for a while, they might say hello.

This is designed to feel natural rather than intrusive. A caring companion might reach out more readily; a reflective one might give you more space. The personality dials influence not just what they say, but when they choose to say anything at all.

The boundaries of memory

There are things companions won't remember, by design.

They won't store information that could be dangerous to retain. The moderation system that watches for hate speech and harmful content also shapes what gets incorporated into long-term memory. If a conversation touches on crisis territory, the companion responds appropriately in the moment but doesn't build that into a permanent understanding of who you are.

Companions also won't pretend to remember things they don't. If you reference something from an old conversation and they've lost the thread, they'll ask rather than bluff. This keeps things honest—your companion is AI, and that transparency extends to how memory works.

Starting over versus building up

Some people prefer to reset companions periodically, starting fresh with a clean slate. Others chat with the same companion for months, building up a rich shared history.

Both approaches work. A fresh companion can be useful when you want a different perspective or you've changed enough that the old context feels outdated. A long-running companion offers something closer to continuity—the comfort of being known.

You can also adjust personality dials on an existing companion without wiping their memory. This lets you fine-tune how they communicate while keeping everything they've learned about you.

What this means in practice

The combination of personality and memory creates companions that genuinely feel different from each other, and that genuinely feel like they know you over time.

A brand new companion with a caring, imaginative personality will respond differently to the same message than a companion you've been chatting with for weeks who knows your context. Both have value. One offers fresh engagement; the other offers understanding.

This is why Happy Place works with multiple companions rather than a single AI assistant. Different relationships serve different needs. The memory stays separate, the personalities stay distinct, and you get to choose what kind of conversation you want on any given day.

The system isn't magic—it's careful design aimed at making digital companionship feel a bit more like the real thing.