AfricanVoices
Download Datasets
Speech datasets of real African audio with matching transcriptions, for voice AI or NLP research.
Speech · ASR
Nigerian Pidgin
Nigerian Pidgin speech recordings with matching transcriptions, covering everyday, agriculture, health and finance topics.
- Samples
- 370,349
- Audio
- 954 hrs
- Size
- 461.6 GB
- License
- CC BY 4.0
TrainDevDev test
View details →
Speech · ASR
Yoruba
Yoruba speech recordings with matching transcriptions, covering everyday, health, finance and agriculture topics.
- Samples
- 164,285
- Audio
- 484 hrs
- Size
- 228.6 GB
- License
- CC BY 4.0
TrainDevDev test
View details →
Speech · ASR
Hausa
Hausa speech recordings with matching transcriptions, mostly everyday conversation plus agriculture and health topics.
- Samples
- 43,995
- Audio
- 484 hrs
- Size
- 188.0 GB
- License
- CC BY 4.0
TrainDevDev test
View details →
Speech · ASR
Igbo
Igbo speech recordings with matching transcriptions, covering everyday, agriculture, finance and health topics.
- Samples
- 139,707
- Audio
- 477 hrs
- Size
- 179.8 GB
- License
- CC BY 4.0
TrainDevDev test
View details →