Version 4.0.0#

Multilingual Speech2Speech Agents are here!#

You can now create your own Multilingual Speech-to-Speech Agent with BAF, supporting the latest OpenAI Speech Models and more. Also works for Luxembourgish! Try it out:

New Features#

  • Added Text2Speech support:

    • Supports State-of-the-art OpenAI Models

    • Supports a variety of HF Models

    • Supports Luxembourgish Piper Model

  • Added more Speech2Text models:

    • Added support for all other HF Models (not just Whisper models)

    • Added support for OpenAI models (through the OpenAI API)

    • Added support for Luxembourgish speech recognition model (LuxASR)

  • Added the audio_language_detection_processor class:

    • Allows you to recognize the language the user speaks and automatically assign the best model

Improvements#

  • Monitoring DB has been updated to store audio messages

Changes#