Techniques for adaptive audio processing during video conferencing are provided. In an example method, a computing system receives an audio stream. The computing system determines, using a machine learning (ML) model, one or more characteristics of the audio stream. The computing system determines, based on the one or more characteristics of the audio stream, an audio configuration operation including one or more instructions to configure, using a programmatic interface, an audio processing component. The computing system executes the audio configuration operation. The computing system processes, by the configured audio processing component, the audio stream. The computing system provides, to a video conference provider, the audio stream.