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

# Episodes

> Episodes are the fundamental unit of ingestion in Ryumem.

Every piece of information starts as an episode.

## Adding Episodes

You can add episodes from various sources:

```python theme={null}
from ryumem import Ryumem

ryumem = Ryumem(server_url="http://localhost:8000")

# Text episode
ryumem.add_episode(
    content="Alice graduated from Stanford in 2018.",
    user_id="user_123",
    session_id="session_abc",
    source="text",
)

# Message episode (conversational)
ryumem.add_episode(
    content="user: Where did you go to school?\nassistant: I went to MIT.",
    user_id="user_123",
    session_id="session_abc",
    source="message",
)

# JSON episode (structured data)
ryumem.add_episode(
    content='{"person": "Bob", "company": "Meta", "role": "Engineer"}',
    user_id="user_123",
    session_id="session_abc",
    source="json",
)
```

## Episode Parameters

| Parameter    | Type | Required | Description                                                  |
| ------------ | ---- | -------- | ------------------------------------------------------------ |
| `content`    | str  | Yes      | The content of the episode                                   |
| `user_id`    | str  | Yes      | User identifier for multi-tenancy                            |
| `session_id` | str  | Yes      | Session identifier for grouping related episodes             |
| `source`     | str  | No       | Episode type: `text`, `message`, or `json` (default: `text`) |
| `kind`       | str  | No       | Episode kind: `query` or `memory` (default: `query`)         |
| `metadata`   | dict | No       | Additional metadata to attach to the episode                 |
| `agent_id`   | str  | No       | Optional agent identifier                                    |

## Episode Kinds

* **query**: Represents user queries or interactions (default)
* **memory**: Represents stored memories or facts extracted from conversations