Data catalogue

African Language Speech Datasets

Commercially licensed Twi, Wolof, and Fon speech data, validated and shipping today with volume scaling now. Every African audio dataset is speaker-verified, AI quality-filtered at capture, and delivered under a commercial license with an SLA.

Available now

Twi, Wolof & Fon speech datasets shipping today

Twi (Akan) Speech Dataset

Ghana · ISO 639: ak

Available · scaling

Natural, code-switched Akan Twi speech for Ghanaian voice and ASR systems.

For: Ghanaian ASR training data & local-language IVR

Modality: Conversational & image-prompted speech
Format: WAV · 16 kHz + timestamped transcripts (JSON)
Licensing: Commercial license + SLA
Quality: >80% inter-annotator agreement
Sourcing: Fair-Trade, native speakers
Delivery: API or S3

Part of our 128-hour speaker-verified Twi, Wolof & Fon corpus fully annotated and growing.

Wolof Speech Dataset

Senegal · ISO 639: wo

Available · scaling

Everyday Wolof speech, including French code-switching, for CPaaS platforms serving Senegal.

For: Wolof IVR Senegal & West African ASR training data

Modality: Conversational & image-prompted speech
Format: WAV · 16 kHz + timestamped transcripts (JSON)
Licensing: Commercial license + SLA
Quality: >80% inter-annotator agreement
Sourcing: Fair-Trade, native speakers
Delivery: API or S3

Part of our 128-hour speaker-verified Twi, Wolof & Fon corpus fully annotated and growing.

Fon Speech Dataset

Benin · ISO 639: fon

Available · scaling

Natural, tonal Fon speech for voice and ASR systems serving Benin and the wider Gbe-speaking belt.

For: Beninese ASR training data & local-language IVR

Modality: Conversational & image-prompted speech
Format: WAV · 16 kHz + timestamped transcripts (JSON)
Licensing: Commercial license + SLA
Quality: >80% inter-annotator agreement
Sourcing: Fair-Trade, native speakers
Delivery: API or S3

Part of our 128-hour speaker-verified Twi, Wolof & Fon corpus fully annotated and growing.

Mission-Based Collection

More African Languages Coming

Need a language that isn't in our catalogue yet? We scope and staff a custom collection mission with our native-speaker network the same image-prompted elicitation, AI quality-filtering and >80% agreement standard as our shipping datasets.

Swahili
Tanzania
Hausa
Nigeria
Amharic
Ethiopia
Yoruba
Nigeria
Zulu
South Africa
Fulfulde
Senegal
Ewe
Togo
Shona
Zimbabwe
Lingala
DR Congo
Igbo
Nigeria
Oromo
Ethiopia
Tigrinya
Ethiopia
Somali
Somalia
Xhosa
South Africa
Luganda
Uganda
Kikuyu
Kenya
Krio
Sierra Leone
Bambara
Mali
Kinyarwanda
Rwanda

Typical lead time: 4–8 weeks from kickoff to first delivery.

Don't see the dataset you need?

Or email us at contact@afriklang.com