> For the complete documentation index, see [llms.txt](https://docs.convai.com/api-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.convai.com/api-docs/plugins-and-integrations/unity-plugin-beta-overview/features/long-term-memory/quick-start.md).

# Quick Start

## Your First Cross-Session Memory

Long-Term Memory requires two things to work: memory must be enabled for the character on the Convai Dashboard, and the SDK must send a stable user identifier with each session. The second part is handled automatically by `DeviceEndUserIdProvider`. This guide covers the minimal steps to get from zero to a character that remembers you.

{% hint style="info" %}
**Prerequisites**

* Your Convai API key is configured in **Tools → Convai → Configuration**.
* A Unity scene with a `ConvaiCharacter` component already set up and working (the character should respond to speech).
  {% endhint %}

## Steps

{% stepper %}
{% step %}
**Enable memory on your character**

Open the [Convai dashboard](https://convai.com) and select your character. Navigate to **Memory → Memory Settings** and enable the **Long-Term Memory** toggle.

This is an opt-in setting — memory is **off by default**. The character will not accumulate or use memories until you enable it here. For a full walkthrough of the dashboard UI, see the [Memory Settings documentation](https://docs.convai.com/api-docs/convai-playground/character-customization/memory#memory-settings).

{% hint style="warning" %}
Enabling memory affects all sessions for that character across all deployments — not just your Unity project. If the character is shared with other team members or projects, coordinate before enabling.
{% endhint %}

<figure><img src="/files/lMbdzXBhmMfypO38r9Of" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}
**Enter Play Mode and start a conversation**

Press **Play** in the Unity Editor. Start a conversation with the character and share a few facts it can remember — your name, your role, or a preference. For example:

> "My name is Alex. I'm a safety officer on the night shift."

Let the character respond, then **stop Play Mode**.
{% endstep %}

{% step %}
**Re-enter Play Mode and verify recall**

Press **Play** again and start a new conversation with the same character. Ask it to recall what it knows:

> "Do you remember who I am?"

If Long-Term Memory is working, the character will reference facts from the previous session without being told again.
{% endstep %}
{% endstepper %}

{% hint style="success" %}
**Expected result:** The character references your name or role without any prompt. The recall will not be verbatim — the AI synthesises stored memories naturally into its replies.
{% endhint %}

## What Just Happened

Behind this seamless experience, four things occurred automatically:

1. **Identity:** The SDK called `DeviceEndUserIdProvider`, which read a stable GUID from `PlayerPrefs` (key: `"convai.end_user_id"`). In the Editor, the same GUID is used every Play Mode session on this machine.
2. **Connect:** That GUID was sent to the Convai server as `end_user_id` when the session started.
3. **Memory recall:** The server resolved the GUID to an internal speaker record, loaded memories for the `speaker_id:character_id` pair, and injected them into the AI's context before the first response.
4. **Memory storage:** At the end of the first session, the backend extracted facts from the conversation (your name, your role) and stored them as `MemoryRecord` entries in the partition.

{% hint style="warning" %}
**Editor identity note:** In the Unity Editor, all Play Mode sessions on the same machine share the same `end_user_id`. This is intentional for stable testing. It does not reflect how player builds behave, where each physical device gets its own identifier. See [End-User Identity](/api-docs/plugins-and-integrations/unity-plugin-beta-overview/features/long-term-memory/end-user-identity.md) for the full picture.
{% endhint %}

## Next Steps

* To understand how the user identifier works and how to replace it with your own: [End-User Identity](/api-docs/plugins-and-integrations/unity-plugin-beta-overview/features/long-term-memory/end-user-identity.md)
* To enable or disable memory programmatically instead of via the dashboard: [Enabling Memory on Characters](/api-docs/plugins-and-integrations/unity-plugin-beta-overview/features/long-term-memory/enabling-memory-on-characters.md)
* To add, list, or delete memory records from code: [Memory Management API](/api-docs/plugins-and-integrations/unity-plugin-beta-overview/features/long-term-memory/memory-management-api.md)
* To view and manage users who have accumulated memories: [End-User Management](/api-docs/plugins-and-integrations/unity-plugin-beta-overview/features/long-term-memory/end-user-management.md)

## Conclusion

With memory enabled on the dashboard and the default identity provider in place, your character will accumulate and recall user-specific facts across every session — entirely automatically. When your project requires user accounts, custom identity, or direct control over memory content, the rest of this section provides the tools to do that.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.convai.com/api-docs/plugins-and-integrations/unity-plugin-beta-overview/features/long-term-memory/quick-start.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.