NeuCodec is an open-source lightweight neural audio codec developed by Neuphonic. It encodes audio at just 0.8 kbps while maintaining near-inaudible reconstruction quality and supports upsampling from 16kHz to 24kHz. It is designed for researchers and developers building speech language models and text-to-speech systems, and includes pre-compressed large datasets such as a compressed version of Emilia-YODAS.
Neucodec is a Voice, TTS & speech product. It focuses on compressing large speech datasets for efficient training of text-to-speech and speech language models without building a custom codec. It is built as an open-source project for AI researchers and developers. Neucodec is open source under the Apache-2.0 license. The product ships for the web, the command line, and API.
Neuphonic builds and maintains Neucodec, and the product first shipped in 2025. Development happens publicly on GitHub with 161 stars and 1 commits in the last 90 days. Key capabilities include low bit-rate compression, 16kHz to 24kHz upsampling, and pre-encoded datasets.
Latest indexed changes and source events
neuphonic/neucodec verified by the PulseGate indexer
Other apps tracked under the same category.
