East Asia · South Korea
Korean conversational speech dataset
500 hours of spontaneous Korean conversations recorded in South Korea across 50 vetted native speakers. Dual-channel audio with time-aligned transcripts, speaker metadata (age, gender, region, device), and full chain-of-consent documentation. Licensed for commercial ASR, TTS, and voice AI training.
Dataset specifications
| Language | Korean |
|---|---|
| Recorded in | South Korea |
| Speech type | Spontaneous two-person conversations |
| Volume | 500 hours |
| Speakers | 50 vetted native speakers |
| Audio | Dual-channel (one channel per speaker) |
| Transcripts | Time-aligned, human-verified |
| Metadata | Age, gender, region, recording device |
| License | Commercial (training & evaluation) |
| Price | $95 per hour of audio |
Need more hours, another dialect or your own recording setup? We collect speech data to spec.
Built for
Training and evaluating speech models that must handle real Korean speech:
- Automatic speech recognition (ASR / STT)
- Text-to-speech & voice synthesis prosody
- Voice agents and IVR for South Korea
- Speaker diarization & turn-taking models
- Speech LLM / audio-understanding fine-tuning
Why conversational data?
Models trained on read speech consistently underperform on spontaneous audio. Disfluencies, overlap, and natural speech rates aren’t in the training distribution. This dataset closes that gap for Korean. Read more in our guide to conversational vs. scripted speech data.
Free samples
Audit the Korean data yourself
Share your company email and we’ll email you representative audio, matching transcripts, and the metadata schema. Check audio quality, transcript accuracy, and speaker diversity before you spend a dollar.
Public alternatives
Public Korean speech data vs. this dataset
| Source | What it is | License | Commercial training |
|---|---|---|---|
| AI Hub Korean Speech (한국어 음성) (opens in a new tab)AI Hub (NIA, South Korea) | About 1,000 hours of free two-person Korean conversation on everyday topics from about 2,000 speakers, recorded in a quiet environment. | AI Hub usage policy: AI training for commercial or non-commercial R&D; open to applicants with Korean nationality only; overseas users need a separate agreement with NIA | Unclear |
| Zeroth-Korean (SLR40) (opens in a new tab)Zeroth Project (Atlas Guide and Gridspace), via OpenSLR | 51.6 hours of transcribed Korean audio for training: 3,000 sentences recorded by 105 people. | CC BY 4.0 | Allowed |
| Common Voice Scripted Speech (Korean) (opens in a new tab)Mozilla Common Voice (Mozilla Data Collective) | Read speech recordings in Korean from Mozilla's Common Voice project. | CC0 1.0 | Allowed |
| FLEURS (Korean) (opens in a new tab)Google | Read sentences from the FLoRes translation benchmark, recorded in 102 languages. | CC BY 4.0 | Allowed |
Overseas users need a separate NIA agreement, since AI Hub's conversations are open to Korean nationals only. Zeroth, Common Voice and FLEURS are read; ours is dual-channel, commercially licensed.
Licenses as published by each source, checked October 2026. Check the current terms before use.
FAQ
Frequently asked questions
What does the Korean speech dataset include?
500 hours of spontaneous two-person conversations from 50 vetted native speakers recorded in South Korea, delivered as dual-channel audio with time-aligned transcripts, speaker metadata (age, gender, region, device), and consent documentation.
What makes Korean hard for speech recognition?
Korean sounds change at word boundaries in ways the Hangul spelling doesn't show, and particles and endings attach to words, so word forms multiply. Speech levels shift between formal and casual depending on who's talking, and conversation is full of English loanwords.
Is there free Korean speech data?
Yes, Zeroth-Korean, Common Voice and FLEURS are free for commercial use. AI Hub Korean Speech has unclear terms. This dataset is spontaneous two-person conversation recorded in South Korea, with speaker metadata, one channel per speaker and signed commercial-use consent.
Can I use this Korean dataset commercially?
Yes. The dataset is licensed for commercial use, including training ASR, TTS, and other voice-AI models. Every speaker has signed explicit commercial-use consent, and we provide the full consent chain for your legal review.
How much does the Korean speech dataset cost?
Pricing is $95 per hour of audio. Volume discounts are available for full-dataset and multi-language purchases. Contact us for a quote.
Can I evaluate samples before buying?
Yes. Request samples with your company email, and we'll email you representative audio files and matching transcripts so your team can audit quality before committing.
$70/hr