Africa · South Africa
English (South African) conversational speech dataset
2,000 hours of spontaneous English (South African) conversations recorded in South Africa across 200 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 | English (South African) |
|---|---|
| Recorded in | South Africa |
| Speech type | Spontaneous two-person conversations |
| Volume | 2,000 hours |
| Speakers | 200 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 | $60 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 English (South African) speech:
- Automatic speech recognition (ASR / STT)
- Text-to-speech & voice synthesis prosody
- Voice agents and IVR for South Africa
- 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 English (South African). Read more in our guide to conversational vs. scripted speech data.
Free samples
Audit the English (South African) 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 South African English speech data vs. this dataset
| Source | What it is | License | Commercial training |
|---|---|---|---|
| NCHLT English Speech Corpus (opens in a new tab)Meraka Institute (CSIR) and North-West University (via SADiLaR) | Orthographically transcribed English speech read from short prompts, part of the NCHLT corpus of South Africa's eleven official languages. | CC BY 3.0 | Allowed |
| AfriSpeech-200 (opens in a new tab)Intron Health | African-accented English read aloud, clinical and general domain, from speakers in several countries including South Africa. | CC BY-NC-SA 4.0 | Not allowed |
| Mozilla Common Voice English (Scripted and Spontaneous Speech) (opens in a new tab)Mozilla Common Voice (via Mozilla Data Collective) | English sentences read by volunteers worldwide, not South Africa-specific; a Spontaneous Speech release adds unscripted answers. | CC0 1.0 | Allowed |
NCHLT and AfriSpeech-200 are prompted or read speech, and AfriSpeech-200 is non-commercial; this dataset is spontaneous two-person conversation recorded in South Africa, licensed for commercial use.
Licenses as published by each source, checked October 2026. Check the current terms before use.
FAQ
Frequently asked questions
What does the English (South African) speech dataset include?
2,000 hours of spontaneous two-person conversations from 200 vetted native speakers recorded in South Africa, delivered as dual-channel audio with time-aligned transcripts, speaker metadata (age, gender, region, device), and consent documentation.
What makes South African English hard for speech recognition?
South African English is shaped by contact with Afrikaans, isiZulu, isiXhosa and other local languages. You hear it in vowel quality, intonation and words like "braai" or "robot" for a traffic light. Speakers also switch into other South African languages mid-conversation, which global English models rarely see in training.
Is there free South African English speech data?
Yes, the NCHLT English Speech Corpus and Mozilla Common Voice English are free for commercial use. AfriSpeech-200 is non-commercial only. This dataset is spontaneous two-person conversation recorded in South Africa, with speaker metadata, one channel per speaker and signed commercial-use consent.
Can I use this English (South African) 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 English (South African) speech dataset cost?
Pricing is $60 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.
$90/hr