> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.jambonz.org/guides/features/conferencing-coach-mode/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.jambonz.org/_mcp/server.

# Conferencing "coach" mode

> Use coach mode in jambonz conferences to whisper to one participant without other callers hearing. Setup and configuration.

The conferencing feature supports a coaching feature whereby designated participants can hear the
full conference audio but their own audio is only heard by a subset of the conference participants.

The classic example of this is a manager listening in on a call between an agent and a customer
and being able to "whisper" to the agent without the customer hearing.
This feature is not limited to a single participant receiving the audio, it could be any subset of the
conference participants.

## Enabling coach mode

To enable coach mode, you must assign a **member tag** to the participants that you may want to selectively send the audio to.

For instance, when an agent joins the conference you might assign them a memberTag of `agent`:

```js
app.conference({
  name: 'support-conference-23',
  startConferenceOnEnter: true,
  endConferenceOnExit: true,
  memberTag: 'agent'
})
```

Later, when the manager joins, you could use the `speakOnlyTo` tag to indicate their audio input should
only be heard by agents:

```js
app.conference({
  name: 'support-conference-23',
  speakOnlyTo: 'agent'
})
```

## Dynamically modifying flows

During the conference, you can dynamically change these settings,
and the conference audio flows will be immediately updated accordingly.
This can be done using the REST API as show below:

```bash maxLines=35
# change a participant's member tag value
curl --location -g --request PUT '{{base_url}}/v1/Accounts/{{account_sid}}/Calls/{{call_sid}}' \
--data '{
    "conferenceParticipantAction": {
      "action": "tag",
      "tag": "customer"
    }
}'

# remove a participant's member tag
curl --location -g --request PUT '{{base_url}}/v1/Accounts/{{account_sid}}/Calls/{{call_sid}}' \
--data '{
    "conferenceParticipantAction": {
      "action": "untag"
    }
}'

# make manager's audio heard by everyone
curl --location -g --request PUT '{{base_url}}/v1/Accounts/{{account_sid}}/Calls/{{call_sid}}' \
--data '{
    "conferenceParticipantAction": {
      "action": "uncoach"
    }
}'

# put a participant into coach mode
curl --location -g --request PUT '{{base_url}}/v1/Accounts/{{account_sid}}/Calls/{{call_sid}}' \
--data '{
    "conferenceParticipantAction": {
      "action": "coach",
      "tag": "agent"
    }
}'
```

The same can be accomplished via the websocket api:

```js
// change a participant's member tag
session.injectCommand('conf:participant-action', {
  action: 'tag',
  tag: 'customer'
});

// remove a particpant's member tag
session.injectCommand('conf:participant-action', {
  action: 'untag'
});

// make manager's audio heard by everyone
session.injectCommand('conf:participant-action', {
  action: 'uncoach'
});

// put a participant into coach mode
session.injectCommand('conf:participant-action', {
  action: 'coach', tag: 'agent'
});
```

> Support for selectively sending audio to a subset of conference participants