LiveKit Transport Provider

The LiveKit provider implements bidirectional audio transport through LiveKit rooms. LiveKit is treated as a transport layer — it provides WebRTC connectivity while Beluga handles all STT, LLM, and TTS processing through the frame-based pipeline.

Choose LiveKit when you need a self-hostable, open-source WebRTC infrastructure for production voice deployments. LiveKit provides room management, access token authentication, and scalable media routing. Beluga treats LiveKit as a pure transport layer, handling all voice processing in the frame-based pipeline. For a managed alternative, consider Daily.

Installation

go get github.com/lookatitude/beluga-ai/voice/transport/providers/livekit

Configuration

Field	Required	Default	Description
URL	Yes	—	LiveKit server URL (wss://...)
Token	Yes	—	LiveKit access token
SampleRate	No	16000	Audio sample rate in Hz
Channels	No	1	Audio channels (1=mono, 2=stereo)

Extra configuration:

Key	Type	Description
room	string	LiveKit room name to join

Basic Usage

package main

import (
    "context"
    "fmt"
    "log"

    "github.com/lookatitude/beluga-ai/voice/transport"
    _ "github.com/lookatitude/beluga-ai/voice/transport/providers/livekit"
)

func main() {
    t, err := transport.New("livekit", transport.Config{
        URL:   "wss://myapp.livekit.cloud",
        Token: "your-livekit-token",
        Extra: map[string]any{
            "room": "voice-room",
        },
    })
    if err != nil {
        log.Fatal(err)
    }
    defer t.Close()

    // Receive incoming audio frames
    frames, err := t.Recv(context.Background())
    if err != nil {
        log.Fatal(err)
    }

    for frame := range frames {
        fmt.Printf("Received %d bytes of audio\n", len(frame.Data))
    }
}

Advanced Features

Sending Audio

Write processed audio back to the LiveKit room using voice.NewAudioFrame:

frame := voice.NewAudioFrame(processedPCM, 16000)
err := t.Send(ctx, frame)
if err != nil {
    log.Printf("send failed: %v", err)
}

Room Configuration

Pass the room name through the Extra map:

t, err := transport.New("livekit", transport.Config{
    URL:   "wss://myapp.livekit.cloud",
    Token: token,
    Extra: map[string]any{
        "room": "my-voice-room",
    },
})

Voice Pipeline Integration

Wrap the transport for use with Beluga’s voice pipeline:

import "github.com/lookatitude/beluga-ai/voice/transport"

audioTransport, err := transport.New("livekit", cfg)
if err != nil {
    log.Fatal(err)
}

voiceTransport := &transport.AsVoiceTransport{T: audioTransport}
// Use voiceTransport with the voice pipeline

Direct Construction

import "github.com/lookatitude/beluga-ai/voice/transport/providers/livekit"

t, err := livekit.New(transport.Config{
    URL:        "wss://myapp.livekit.cloud",
    Token:      "your-livekit-token",
    SampleRate: 16000,
    Channels:   1,
    Extra:      map[string]any{"room": "voice-room"},
})

Error Handling

t, err := transport.New("livekit", transport.Config{
    URL: "wss://myapp.livekit.cloud",
    // Missing Token
})
if err != nil {
    // "livekit: Token is required"
    log.Fatal(err)
}

Both URL and Token are required. The transport returns errors from Recv and Send if Close has already been called.

Refer to the LiveKit documentation for server setup, room management, and token generation.