> ## Documentation Index
> Fetch the complete documentation index at: https://docs.wearecorgi.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# How Token remembers

> Token keeps the context of each conversation, separately and privately.

Token remembers the context of a conversation, so you do not have to repeat
yourself. Ask a follow-up and it already knows what you were talking about.

## Memory is per conversation

Each chat has its own memory, kept separate from every other chat. What you
discuss in one channel does not carry into another, and a direct message stays
private to that conversation. It is the same boundary that keeps teammates' work
separate. See [Teams and accounts](/teams).

<Check>
  Because memory is per conversation, you can keep a long-running thread going and
  Token will remember the earlier context the next time you message in that same
  chat.
</Check>

## Keeping things in mind

If you tell Token something to keep in mind for a conversation, like a preference
or a recurring detail, it can hold on to it for that chat. Just say so in plain
language.

> Keep in mind that our launch date is March 3.

Later in the same conversation, Token can use that without you repeating it.

## Long conversations

As a conversation gets very long, Token tidies up older context on its own so it
can keep going. You do not need to do anything; it happens automatically.

<Note>
  Memory lives with the conversation. If you want Token to do something on a
  regular basis, a [routine](/routines) is usually a better fit than
  relying on memory alone.
</Note>