ultemica/wuthering-waves-voices

Wuthering Waves — Voice Lines (Multi-Language)

An archive of character voice data extracted from Wuthering Waves (鳴潮), repackaged as Parquet shards per audio language.

Dataset Summary

Field	Value
Game	Wuthering Waves (鳴潮)
Publisher	Kuro Games (Kuro Game Studio)
Languages	中文 (zh), 日本語 (ja), English (en), 한국어 (ko)
Game version	2.2
Source format	WAV + sidecar transcripts (.lab / .txt)
Distribution format	Apache Parquet (zstd), ~500 MiB audio per shard
Extraction tool	AI-Hobbyist / WutheringWaves_Datasets (ModelScope mirror)

Directory Layout (HF repo)

data/
├── zh/train-NNNNN-of-NNNNN.parquet
├── ja/train-NNNNN-of-NNNNN.parquet
├── en/train-NNNNN-of-NNNNN.parquet
└── ko/train-NNNNN-of-NNNNN.parquet

Each language is a separate config. HF auto-detects the layout — load by passing the language code as the second argument to load_dataset.

Loading

from datasets import load_dataset

ds = load_dataset("ultemica/wuthering-waves-voices", "ja", split="train", streaming=True)
sample = next(iter(ds))
print(sample["transcription"], sample["speaker"])

Schema

Field	Type	Notes
audio	{ bytes, path }	WAV bytes + original filename
transcription	string | null	Sidecar .lab / .txt text if available
speaker	string	Top-level folder in the source archive (character / NPC)
emotion	string | null	Reserved; currently null
language	string	Matches the config name
source_path	string	Original WAV path inside the 7z archive

Intended Use

• Training / evaluation data for ASR (automatic speech recognition) models
• Reference audio for TTS (text-to-speech) systems
• Speaker diarization / speaker recognition tasks
• Research on multilingual anime-style voice acting

Note: This dataset is intended for research and non-commercial use only.

Data Source & Extraction

The original assets were obtained from the official Wuthering Waves client distributed by Kuro Games, then decrypted and packaged by the AI-Hobbyist community on ModelScope. This repository repackages those archives into language-keyed Parquet shards suitable for streaming via the Hugging Face datasets library.

• GitHub: https://github.com/AI-Hobbyist/WutheringWaves_Datasets
• ModelScope: https://modelscope.cn/datasets/aihobbyist/WutheringWaves_Dataset

License & Copyright

All assets are © Kuro Games. Wuthering Waves is a copyrighted work of Kuro Games (Kuro Game Studio). This repository is a fan-made archive intended solely for research and educational purposes; no commercial use is permitted. We will respond promptly to any takedown request from the rights holders.

• Copyright of the original assets belongs entirely to Kuro Games.
• Commercial use of any model or derivative work trained on this dataset is strongly discouraged.
• If you redistribute this dataset, please include this notice.

Citation

@misc{wuthering_waves_voice_2026,
  title  = {Wuthering Waves Voice Lines (Game Version 2.2)},
  author = {ultemica},
  year   = {2026},
  howpublished = {\url{https://huggingface.co/datasets/ultemica/wuthering-waves-voices}},
  note   = {Extracted assets © Kuro Games}
}