xAI STT

The xAI STT provider enables your agent to use xAI's speech-to-text models for converting audio input to text.

Installation

Install the xAI-enabled VideoSDK Agents package:

pip install "videosdk-plugins-xai"

Authentication

The xAI plugin requires an xAI API key.

Set XAI_API_KEY in your .env file.

Importing

from videosdk.plugins.xai import XAISTT

Example Usage

from videosdk.plugins.xai import XAISTT
from videosdk.agents import Pipeline

# Initialize the xAI STT model
stt = XAISTT(
    # When XAI_API_KEY is set in .env - DON'T pass api_key parameter
    api_key="your-xai-api-key",
    language="en",
    sample_rate=48000,
)

# Add stt to pipeline
pipeline = Pipeline(stt=stt)

note

When using .env file for credentials, don't pass them as arguments to model instances or context objects. The SDK automatically reads environment variables, so omit api_key, videosdk_auth, and other credential parameters from your code.

Configuration Options

• api_key: Your xAI API key (required, can also be set via environment variable)
• sample_rate: Audio sample rate in Hz. xAI accepts 8000/16000/22050/24000/44100/48000. Default: 48000
• encoding: Raw audio encoding. One of "pcm" (signed 16-bit LE), "mulaw", "alaw". Default: "pcm"
• interim_results: Emit partial transcripts as they arrive. Default: True
• endpointing: Silence duration (ms) before xAI fires utterance-final. Range 0–5000. Default: 50
• language: BCP-47 language code (e.g. "en", "fr"). Default: "en"
• diarize: When true, each word in the response includes a speaker field. Default: False
• multichannel: When true, transcribes each input channel independently. Default: False
• channels: Number of input channels (only relevant with multichannel=True). Default: 1
• base_url: WebSocket endpoint URL. Default: "wss://api.x.ai/v1/stt"

Additional Resources

The following resources provide more information about using xAI (Grok) with the VideoSDK Agents SDK.

• xAI docs: xAI STT API documentation.

SDK Reference

GitHub Repository

Python Package