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

# Control What Maya Remembers About You

> Manage Kolmena's long-term memory for Maya and your agents. Configure memory settings, view what is remembered, and edit or delete individual memories.

Kolmena builds a long-term memory from your conversations so Maya and your agents can stay context-aware across chats. You control whether memory is on, who it applies to, and how it is shared. You can also view, edit, or delete anything that has been remembered.

## Memory Settings

Open **Settings** and select the **Memory** tab to configure how Kolmena remembers.

<Steps>
  <Step title="Toggle memory on or off">
    Choose whether Kolmena builds a long-term memory at all. When memory is off, agents rely only on the current conversation context.
  </Step>

  <Step title="Choose who remembers">
    Select **Maya only** if you want memory to help your personal assistant but not other agents. Select **All agents** to let every agent in your hive draw on shared context.
  </Step>

  <Step title="Pick a processing model">
    Choose which AI model reads your conversations to decide what is worth remembering. A stronger model extracts more nuanced detail and costs more per message. Changing the model only affects new memories; existing ones stay as they are.
  </Step>

  <Step title="Set visibility">
    Choose **Shared brain** so every agent can draw on the same memory pool. Choose **Private per agent** so each agent's new memories stay separate. Note: switching to private does not pull back memories already shared across agents.
  </Step>
</Steps>

## View Your Memories

In the Memory settings page, scroll to the **Memories** section. You will see a searchable table of everything Kolmena has learned about you.

| Column | What it shows |
| - | - |
| **Memory** | A short summary of the fact or experience. |
| **Type** | Whether it is a **Fact** or an **Experience**. |
| **Source** | Where it came from: chat, Telegram, jobs, swarm, API, or saved by an agent. |
| **Agent** | Which agent learned it, or **Shared** if it lives in the shared brain. |
| **Date learned** | When the memory was created. |
| **Provenance** | Whether it is shared across agents, private to one agent, or limited to a single conversation. |

Use the search box to find specific memories by keyword.

## Edit a Memory

If Kolmena remembered something incorrectly, you can correct it.

<Steps>
  <Step title="Find the memory">
    Search or scroll to the memory you want to change.
  </Step>

  <Step title="Click Edit memory">
    Click the edit action on the row.
  </Step>

  <Step title="Correct it">
    Type what Kolmena should remember instead, then save.
  </Step>
</Steps>

## Delete a Memory

You can remove individual memories or clear everything at once.

* **Delete one memory**: Click the delete action on a single row.
* **Clear all memory**: Look for the option to forget all memories. This removes everything from Hindsight and cannot be undone.

<Warning>
  Deleting memories is permanent. If you clear all memory, agents will start fresh and may need you to re-explain preferences or context.
</Warning>

## Ask Your Brain

In the Memory settings, use the **What Kolmena understands about me** panel to ask natural-language questions about what is stored. For example, ask "What projects am I working on?" or "What tools do I prefer?" Kolmena searches your memory and answers based on what it knows.

<Note>
  If the profile panel is empty, it means Kolmena has not learned enough yet. Keep chatting and it will fill in over time.
</Note>

## Next Steps

<CardGroup cols={2}>
  <Card title="Chat with Maya" color="#FE4F32" icon="message-dots" href="/maya/chat-with-maya">
    Conversations are the main source of memories.
  </Card>

  <Card title="Outputs" color="#FE4F32" icon="files" href="/maya/outputs-and-deliverables">
    See what your agents produce from that context.
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.