Core Viewpoint - Meituan's LongCat team has officially open-sourced a dedicated audio codec solution named LongCat-Audio-Codec, which provides a comprehensive toolchain for token generation and restoration, facilitating efficient audio processing for Speech LLMs [1] Group 1 - The LongCat-Audio-Codec includes a one-stop Tokenizer and DeTokenizer toolchain [1] - The core functionality of the codec is to map raw audio signals into a token sequence that is semantically and acoustically parallel [1] - The solution enables high-quality audio reconstruction through its decoding module, supporting full-link audio processing from signal input to output [1]