We've added new customization options for our transcription service. You can now specify the language and model used for transcription.

New Customization Options

We've added two new optional parameters to our bot deployment API:

• language: Specify the language for transcription (defaults to 'multi')
• model: Choose the Deepgram model for transcription (defaults to 'nova-3')

Available Options

We use Deepgram for transcription. According to the Deepgram documentation, here are the key models:

• nova-3: The highest performing model, recommended for most use cases, especially audio with multiple languages, background noise, crosstalk and far field audio
• nova-2: Recommended for use cases with non-English transcription and filler word identification

How to Use These Features

You can specify these options when deploying a bot:

from chatterbox_io import ChatterBox

client = ChatterBox(authorization_token="your_api_token")

# Deploy a bot with custom language and model settings
session = await client.send_bot(
    platform="zoom",
    meeting_id="123456789",
    language="es",  # Spanish
    model="nova-3"  # Latest model
)

Best Practices

1. Use multi for meetings with multiple languages
2. Specify the exact language code for single-language meetings
3. Use nova-3 for most use cases, as it offers the best overall performance
4. Consider nova-2 for non-English transcription needs

For a complete list of available models and supported languages, visit the Deepgram documentation.