Dataset catalog
License conversational speech datasets
SpeechData.ai lists conversational speech datasets across 60 languages and dialects. Every dataset ships with dual-channel audio, time-aligned transcripts, speaker metadata, and documented consent, licensed for commercial ASR, TTS, and voice-AI training at $60 to $95 per hour. Can’t find your language? We collect to spec.
Need Arabic speech? Compare Modern Standard and regional Arabic datasets before choosing samples.
Showing 3 of 60 datasets
$65/hr