> ## Documentation Index
> Fetch the complete documentation index at: https://docs.60db.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# 60db Extensions

> Smart Memory, Knowledge Collections, intent filler speech and multi-language transcription in 60db Realtime sessions

60db adds four fields to the OpenAI session. OpenAI ignores them, so the same session config runs on both services. Add them to get memory, knowledge answers, spoken acknowledgements and multi-language transcription.

| Field | What it does |
| - | - |
| `memory` | Recalls what this user said in past sessions, and saves the session at close |
| `knowledge` | Answers from your Memory → Collections |
| `intent` | Speaks a short acknowledgement while the reply is written, and passes hints to the LLM |
| `audio.input.transcription.language` | Transcribes several languages in one session (array form) |

## Smart Memory

Turn on `memory` to recall past conversations with the same user.

```json theme={null}
{
  "type": "session.update",
  "session": {
    "type": "realtime",
    "memory": true
  }
}
```

Full form:

```json theme={null}
{
  "memory": { "enabled": true, "persist": true }
}
```

* **Each turn:** the server searches the user's memory and adds the relevant context to the prompt. The model uses it without reading out the memory system.
* **At close:** the session transcript is saved to the user's memory, so the next session can use it.
* **Permission:** the API key needs the **memory** permission. Otherwise the session continues without memory.
* **Scope:** memory is per user. Memories show in the Memory pages and the [Memory API](https://docs.60db.ai/features/memory.md).

> [!NOTE]
>
> Each memory or knowledge lookup that adds context to a reply costs \$0.0003. A turn with nothing relevant found is not charged.

## Knowledge collections

Point a session at Memory collections so answers come from your documents.

```json theme={null}
{
  "knowledge": { "collections": ["company_handbook"] }
}
```

* Up to **5** collection IDs per session.
* Create and upload collections in **Memory → Collections**, or with the [Memory API](https://docs.60db.ai/features/memory.md).

## Intent filler and hints

`intent` lets a small decision model choose a fitting spoken acknowledgement while the main reply is written. This hides latency on questions that take a moment.

```json theme={null}
{
  "intent": { "filler": true, "hints": false }
}
```

| Option | Default | Effect |
| - | - | - |
| `filler` | `false` | Speaks a short acknowledgement, for example "जी, एक सेकंड।" |
| `hints` | `false` | Passes a note to the LLM, for example that the caller wants a human |

* No filler is spoken for greetings or goodbyes.

## Multi-language transcription

Pass an array of language codes to transcribe a mixed-language conversation:

```json theme={null}
{
  "audio": {
    "input": {
      "transcription": { "language": ["en", "hi"] }
    }
  }
}
```

* A single string such as `"en"` is also accepted.
* When `language` is omitted, the default is English and Hindi.
* The reply uses the user's language, in its native script.

## Compatibility

| Field | OpenAI | 60db |
| - | - | - |
| `memory`, `knowledge`, `intent` | Ignored | Applied |
| `transcription.language` as array | Not standard | Applied |

## Related

* [Audio and turn detection](https://docs.60db.ai/realtime/audio-and-turn-detection.md)
* [Memory & RAG](https://docs.60db.ai/features/memory.md)


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