Your First Realtime AI Call

Install mod_realtime_ai

Install the module package for your FreeSWITCH platform, then make sure mod_realtime_ai is loaded by FreeSWITCH.

Create the FreeSWITCH Configuration

Create conf/autoload_configs/realtime_ai.conf.xml and define at least one realtime AI profile. This example creates an OpenAI profile named sales.

<configuration name="realtime_ai.conf" description="Realtime AI">
    <profiles>
        <profile name="sales">
            <param name="provider" value="openai"/>
            <param name="api-key" value=""/>
            <param name="config" value="sales.json"/>
            <param name="model" value="gpt-realtime-2.1"/>
        </profile>
    </profiles>
</configuration>

Add your provider API key to api-key. A single configuration file can contain as many provider profiles as you need.

Create the Provider Configuration

Create conf/realtime_ai/sales.json. The JSON uses the native configuration format required by the selected provider.

{
  "type": "session.update",
  "session": {
    "type": "realtime",
    "instructions": "You are a helpful voice assistant.",
    "output_modalities": ["audio"],
    "audio": {
      "input": {
        "format": {
          "type": "audio/pcma"
        },
        "turn_detection": {
          "type": "server_vad",
          "threshold": 0.7,
          "prefix_padding_ms": 300,
          "silence_duration_ms": 500,
          "create_response": true,
          "interrupt_response": true
        }
      },
      "output": {
        "format": {
          "type": "audio/pcma"
        },
        "voice": "marin"
      }
    }
  }
}

Connect the Call

Start realtime AI on an active FreeSWITCH channel using the profile name.

Your Call Is Connected

mod_realtime_ai now handles realtime audio between the FreeSWITCH call and the selected AI provider.

The same profile-based workflow is used with OpenAI, Google Gemini, ElevenLabs and Azure Voice Live.