xAI Realtime Integration

xAI's Grok Realtime API provides speech-to-speech conversation with <700ms latency. Unlike traditional voice AI (STT → LLM → TTS), Grok processes audio directly in a single model.

Setup

1. Add xAI API Key

Navigate to Integrations → API Keys tab and add your xAI API key:

curl -X POST https://api.flireo.com/api/v1/byok 
  -H "Authorization: Bearer YOUR_API_KEY" 
  -H "Content-Type: application/json" 
  -d '{
    "provider": "xai",
    "api_key": "xai-..."
  }'

2. Configure Agent

When creating or editing an agent with xAI configured:

• Provider: Select "xAI Realtime"
• Model: grok-realtime-v1
• Voice: ara (or other available voices)

Note: When using xAI Realtime, separate STT/TTS providers are ignored.

Backend Integration

Your backend agent will use:

from livekit.plugins import xai

session = AgentSession(
    llm=xai.realtime.RealtimeModel(
        api_key=xai_api_key,  # From Vault
        voice="ara",
    )
)

Pricing

xAI Realtime uses BYOK pricing:

• €0,07/minute when using your xAI API key
• Direct billing to your xAI account
• No markup on xAI API usage

Differences from Traditional Mode

Feature	Traditional (STT+LLM+TTS)	xAI Realtime
Latency	~1-2 seconds	<700ms
Providers	3 separate (Deepgram + OpenAI + ElevenLabs)	Single (xAI)
Voice Quality	Depends on TTS provider	Native to model
Custom Tools	Supported via llm_config.tools	Check xAI docs for support
BYOK Keys Required	3 keys (STT, LLM, TTS)	1 key (xAI)

Limitations

• Custom system prompts may work differently than OpenAI
• Tool calling support depends on xAI API capabilities
• Voice selection limited to xAI's available voices

API Reference

See BYOK API Reference for managing xAI API keys.