> 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/web-plugins/convai-web-sdk/long-term-memory.md).

# Long Term Memory

### Enable memory

This is a character based feature. Go to you convai dashboard -> Character -> Memory tab -> Memory Settings -> Enable Long Term Memory

<figure><img src="/files/76nWR8ZR9siyIbAa8kFL" alt=""><figcaption></figcaption></figure>

Pass `endUserId` when connecting:

```ts
const client = useConvaiClient({
  apiKey: 'YOUR_API_KEY',
  characterId: 'YOUR_CHARACTER_ID',
  endUserId: 'a1b2c3d4-...',  // any string — (UUID or email) preferred, e.g. 'user@example.com'
});
```

`client.memoryManager` becomes available after a successful connection.

### Access the manager

```ts
const memory = client.memoryManager;

if (!memory) {
  // endUserId was not provided, or client is not yet connected
  return;
}
```

***

### List memories

```ts
const result = await memory.listMemories({ page: 1, pageSize: 50 });

console.log(`Total: ${result.total_count}`);
console.log(`Has more: ${result.has_more}`);

result.memories.forEach(m => {
  console.log(`[${m.id}] ${m.memory}`);
  // m.created_at, m.updated_at (ISO timestamps)
});

// Paginate if needed
if (result.has_more) {
  const page2 = await memory.listMemories({ page: 2, pageSize: 50 });
}
```

#### Parameters

| Field      | Type     | Default | Range  |
| ---------- | -------- | ------- | ------ |
| `page`     | `number` | `1`     | 1–1000 |
| `pageSize` | `number` | `50`    | 1–100  |

***

### Add memories

Pass one or more strings to add as memories:

```ts
const result = await memory.addMemories([
  'User prefers dark mode UI.',
  'User is learning Spanish.',
  'User plays guitar as a hobby.',
]);

result.memories.forEach(m => {
  console.log(`Added: ${m.id} → ${m.memory}`);
});
```

The character will use these in future conversations automatically.

***

### Get a single memory

```ts
const m = await memory.getMemory('f4cbdb08-7062-4f3e-8eb2-9f5c80dfe64c');

console.log(m.memory);      // "User prefers dark mode UI."
console.log(m.created_at);  // ISO timestamp
console.log(m.updated_at);  // ISO timestamp
```

***

### Delete a memory

```ts
const result = await memory.deleteMemory('f4cbdb08-7062-4f3e-8eb2-9f5c80dfe64c');

if (result.deleted) {
  console.log('Deleted:', result.memory_id);
}
```

***

### Delete all memories

Removes all memories for the (character, user) pair. Deletion is asynchronous on the server.

```ts
const result = await memory.deleteAllMemories();
console.log(result.message); // "Memory deletion in progress..."

// Wait briefly and verify
await new Promise(r => setTimeout(r, 2000));
const check = await memory.listMemories();
console.log('Remaining:', check.total_count);
```

***

### Standalone usage

`MemoryManager` can be used independently of the client (e.g., in a backend admin tool):

```ts
import { MemoryManager } from '@convai/web-sdk/core';

const manager = new MemoryManager(
  'YOUR_API_KEY',      // or auth token
  'CHARACTER_ID',
  'END_USER_ID',
);

const memories = await manager.listMemories();
```

***

### Memory object shape

```ts
interface Memory {
  id: string;          // UUID
  memory: string;      // Text content
  created_at: string;  // ISO 8601 timestamp
  updated_at: string;  // ISO 8601 timestamp
}
```

***

### How automatic memory works

When `endUserId` is set, the Convai backend extracts meaningful facts from each conversation and stores them as memories. On future connections with the **same** `endUserId`, these memories are injected into the character's context so it "remembers" the user. This is why the value must be stable and unique per user — a UUID or email address both work well.

You do not need to call any Memory API methods for this to work. The explicit CRUD methods are for reading, seeding, or pruning memories from your application.


---

# 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/web-plugins/convai-web-sdk/long-term-memory.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.