Logo
FROM CALL TO AI IN MINUTES
01
MODULE

Install mod_realtime_ai in FreeSWITCH

02
CONFIGURE

Create a profile for your AI provider

03
CONNECT

OpenAI · Gemini · ElevenLabs · Azure

04
TALK

Realtime bidirectional audio

$ uuid_realtime_ai <uuid> start sales
+OK Success
Native FreeSWITCH Full Duplex Auto Resampling Barge-in
REALTIME AI FOR FREESWITCH
From the developers of mod_audio_stream

Bring Realtime AI
to FreeSWITCH

Native realtime speech-to-speech integration for FreeSWITCH.
Connect to OpenAI, Gemini, ElevenLabs and Azure Voice Live through one module.

Get Started Documentation

Everything You Need for Realtime AI Voice

Native FreeSWITCH integration for realtime speech-to-speech AI — from audio streaming
to provider connectivity and playback.

Native FreeSWITCH Integration

Built directly for FreeSWITCH with simple per-call control through uuid_realtime_ai.

Multiple AI Providers

Connect to OpenAI, Gemini, ElevenLabs and Azure Voice Live through one module.

Full-Duplex Realtime Audio

Stream caller audio and play AI-generated speech back into the call in realtime.

Automatic Audio Resampling

Automatically convert audio between the FreeSWITCH channel and provider-required sample rates.

Server VAD & Barge-In

Use provider-native voice activity detection for responsive, natural realtime conversations.

Profile-Based Configuration

Configure provider, model and session settings once, then start calls using a named profile.

MULTIPLE PROFILES

Configure Once. Choose a Profile per Call.

Define reusable AI profiles for different applications, providers and models.
FreeSWITCH only needs the profile name to start the realtime AI session.

Provider Independent

Use OpenAI, Gemini, ElevenLabs or Azure Voice Live without changing your FreeSWITCH call logic.

Native Provider Configuration

Keep provider-specific session settings in native JSON configuration files instead of translating them into another configuration format.

Per-Call Profile Selection

Create profiles for sales, support or other workflows and select the required configuration when the call starts.

PROFILE CONFIGURATION 
<profiles>
  <profile name="sales">
    <param name="provider" value="openai"/>
    <param name="config" value="sales.json"/>
  </profile>

  <profile name="support">
    <param name="provider" value="elevenlabs"/>
    <param name="config" value="support.json"/>
  </profile>
</profiles>
$ uuid_realtime_ai <uuid> start sales

Building Your Own Voice AI Stack?

Use mod_audio_stream for realtime FreeSWITCH audio streaming over WebSocket and connect calls to your own AI, transcription or media processing backend.

WebSocket Full Duplex RAW Audio G.711 OPUS
Explore mod_audio_stream
SIMPLE PRICING

One License. Everything Included.

Run realtime AI voice directly inside FreeSWITCH with all supported providers.

COMMERCIAL LICENSE

€990 / year

Per FreeSWITCH instance

Start 30-Day Trial
  • OpenAI, Gemini, ElevenLabs & Azure Voice Live
  • Full-duplex realtime audio
  • Automatic audio resampling
  • Server-side VAD
  • Profile-based configuration
  • Updates & support

Need the source code?

Source code licensing is available on request for customers requiring full source access and custom deployment rights. Contact us →

FAQ

Questions Before You Get Started?

Which FreeSWITCH versions and platforms are supported?
mod_realtime_ai supports FreeSWITCH 1.11.x on Debian 12 and Debian 13 installations using the official SignalWire repositories. EL9-compatible distributions using the OKay repository are also supported. With a source code license, the module can also be built for FreeSWITCH 1.10.x and 1.11.x on other Linux distributions where FreeSWITCH itself can be built, including older environments such as CentOS 7.
Is there a free version?
Yes. A free version is available with a limit of 10 concurrent channels. It can be used for both testing and commercial deployments, making it suitable for development and smaller production environments.
What is the 30-day trial for?
The 30-day evaluation version removes the channel limitation and is intended for heavy-load, concurrency, stability and integration testing. It allows you to evaluate the module under realistic production workloads before purchasing a commercial license.
What does the €990/year license include?
The commercial license covers one FreeSWITCH instance and includes all supported realtime AI providers, module features, updates and support during the active subscription period.
Is source code licensing available?
Yes. Source code licensing is available separately on request for customers who require full source access or custom deployment rights. Contact us to discuss your requirements.
Do I need separate accounts with the AI providers?
Yes. mod_realtime_ai provides the native FreeSWITCH integration, while provider accounts, API credentials and usage charges are managed directly with OpenAI, Google, ElevenLabs or Microsoft Azure.
Does the module handle different audio sample rates?
Yes. mod_realtime_ai automatically resamples audio between the FreeSWITCH channel rate and the sample rate required by the selected realtime AI provider.