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

# Understanding Multi-Message Replies on DumGum

When `replyParameters.multipleMessagesPerReply` is enabled (either per request or at the project configuration level), AI replies can be split into multiple shorter messages instead of a single long one.

For example, instead of sending:

> Hey Anthony! How are you this morning? I'm exhausted, didn't sleep well, it's way too hot lately 🥵

The system might send:

> Hey Anthony! How are you this morning?\
> I'm exhausted, didn't sleep well, it's way too hot lately 🥵

This feature is designed to improve readability and make conversations feel more natural and dynamic.

## How it works

* When activated, replies **may** be split. It's not systematic. Several generation strategies and a level of randomness are involved.
* Messages are always sent **in order**.
* By default, messages are sent **instantly**, one after another.
* A new message is sent **only after** the previous one has been successfully delivered.

## Managing message pacing

If you want a more natural pacing (instead of instant bursts), you should manage the timing between messages.
There are two options:

1. **Use our built-in delay system** via `replyParameters.replyTypingDelay` (recommended).
   You can choose between `NORMAL` or `FAST` delays.
2. **Handle the delay yourself** on your side.

When using our delay system:

* Each message is preceded by an `answer.processing` event (so you can show a typing indicator).
* Then a pause is applied.
* Then the message is sent.

Message order is always preserved.
Note that every message sent has its own `answer.created` event.

## What about usage and billing?

Only the first message in a split reply counts toward your usage. The additional messages that follow are free.

So if the AI sends a reply as 3 messages instead of 1, you're only charged for 1: the other 2 don't cost anything extra.

## How to spot free messages in the API

Each `answer.created` event includes fields that let you distinguish the main message from free follow-ups:

* **`sequenceNumber`**: `0` for the main (billable) message, `1`, `2`, `3`… for follow-ups.
* **`parentAnswerId`**: `null` for the main message. For follow-ups, contains the ID of the previous message in the sequence.
* **`cost`**: reflects actual billing. Follow-up messages always have a cost of `"0.0000"`.

A simple rule: if `sequenceNumber > 0`, the message is free.