MOSS-Audio-Tokenizer is an open neural audio codec designed for converting speech audio into compact discrete tokens. It serves as a foundational component for the MOSS TTS family and other speech AI models. The model is distributed on Hugging Face with full weights under an Apache-2.0 license and integrates directly with the Transformers library.
In the Foundation models & chat space, MOSS Audio Tokenizer takes a focused approach. It focuses on converting raw audio into discrete tokens for efficient speech modeling and text-to-speech systems. MOSS Audio Tokenizer is an open-source project aimed at AI researchers and developers. The project is open source (Open Source). MOSS Audio Tokenizer is available on the web and API.
It is developed by OpenMOSS-Team, and the product first shipped in 2025. Among its 4 catalogued features are Feature Extraction, Neural Codec, and Speech Tokenization.
Latest indexed changes and source events
OpenMOSS-Team/MOSS-Audio-Tokenizer verified by the PulseGate indexer
Other apps tracked under the same category.
