mlx-audio

This library addresses the need for efficient speech processing on Apple Silicon by providing text-to-speech (TTS), speech-to-text (STT), and speech-to-speech (STS) capabilities optimized specifically for M series chips. It leverages Apple's MLX framework to deliver fast inference and supports multilingual models, voice cloning, and adjustable speech speed. Its integration includes a Python API, a command-line interface, a Swift package for iOS/macOS, and an OpenAI-compatible REST API, making it versatile for Apple ecosystem applications. With over 6,300 stars and recent commits in March 2026, it shows active maintenance but lacks notable growth history, indicating moderate adoption. Teams focused on Apple hardware and requiring local speech synthesis or recognition will benefit most. It is not suitable if cross-platform support or extensive community models like Mozilla TTS or OpenAI Whisper are priorities, as it is tightly coupled to Apple Silicon and MLX framework.

Choose this library when building speech applications targeting Apple Silicon devices that demand fast, local inference and integration with macOS or iOS. For projects requiring broader platform support or more mature speech-to-text models, alternatives like Whisper or Mozilla TTS should be considered.

It is typically used in products requiring on-device speech synthesis or recognition on Apple hardware, such as voice assistants, accessibility tools, or localized speech interfaces within iOS/macOS apps. Teams building custom voice experiences or interactive applications with Swift integration will find it particularly relevant.

Latest commit: 2026-03-21. Activity data is based on daily RepoPi snapshots of the GitHub repository.