The collection covers English, Kazakh, Kyrgyz, Russian, Tajik, Turkish, Turkmen, Uzbek, and mixed-language speech. Access requests are reviewed manually.

Configurations

audio: utterances with embedded audio.
prompt_specs: structured text and delivery specifications.
clone_pairs: same-speaker reference and target pairs.
voice_profiles: stable speaker and delivery profiles.

The mode field distinguishes freeform, voice_design, and voice_clone. The multilingual_switch split contains one-sentence language-switching material.

Core columns

utt: unique utterance identifier
text: training transcript
spk: speaker or profile identifier
instruct: voice-design instruction when applicable
audio: embedded waveform in the audio configuration

Language fields identify the primary language and every language present in mixed-language rows. Performance metadata includes delivery, emotion, scene, style, event tags, pauses, fillers, disfluencies, and pronunciation guidance.

Snapshot

Split	Prompts	Audio	Clone pairs	Profiles
english	11,948	11,948	600	84
kazakh	11,922	11,921	600	84
kyrgyz	11,894	11,893	600	84
multilingual_switch	12,808	12,758	0	0
russian	11,851	11,851	600	84
tajik	11,889	11,889	600	84
turkish	11,938	11,938	600	84
turkmen	11,907	10,312	59	84
uzbek	14,855	14,854	748	84
Total	111,012	109,364	4,407	672

Loading

Use the datasets library with repository instinct-org/multilingual-tts-voice-dataset, the required configuration, and a named split. Authentication is required after access approval.

Downloads last month: 37

Total file size:

105 GB