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

# Memories

> What the assistant remembers about you across conversations, and how to control it.

**Memories** let the assistant carry useful facts across conversations — your role, a recurring project, a preference you've stated — instead of you repeating yourself every time. Memories are personal: only you see your own.

<Info>
  Memories is an optional organization feature. If you don't see **Memories** in your sidebar or in Settings, ask your administrator.
</Info>

***

## How Memories Get Created

The assistant can save a memory on its own when something worth remembering comes up in a conversation — for example, "the client's name is Acme Corp" or "I prefer bullet points over long paragraphs." You can also add one yourself.

<Steps>
  <Step title="Open Memories">
    Go to **Customize** or your sidebar and open **Memories** to see everything currently saved.
  </Step>

  <Step title="Add one manually">
    Click **Create Memory**, then fill in a **Key** and a **Value** — for example, key `preferred_language`, value `French`.
  </Step>
</Steps>

## Viewing, Editing, and Deleting Memories

Every saved memory appears in the **Memories** view with its key and value. From there you can:

* Edit an existing memory's value
* Delete a memory you no longer want kept
* Filter the list to find a specific memory

<Tip>
  Memory keys must use lowercase letters and underscores only (for example `favorite_tools`, not `Favorite Tools`).
</Tip>

## Scoping Memory to a Project

By default, memories apply across all your conversations (**User** scope). If [Projects](/en/improved-usage/projects) is enabled for your organization, you can instead scope a memory to **This project**, so it only applies while you're working inside that project.

## Storage Limits

Your organization sets a total size for your saved memories (a token and character budget). If you're near the limit, WonkaChat tells you when creating or updating a memory would exceed it — delete something you no longer need to make room.

## Turning Memories Off

If you'd rather the assistant not use what it has saved, go to **Settings → Memories** and turn off **Reference saved memories**. This stops the assistant from using your saved memories in its responses without deleting them.

<Note>
  Your administrator can also disable memory creation, reading, updating, or opting out entirely at the organization level. If any of these controls are missing for you, that's likely why.
</Note>

***

## For Admins

Organization owners control memory permissions from **Settings → Organization**: allow creating, reading, and updating memories, and whether users are allowed to opt out. Memories can be disabled for the whole organization at any time.