Spitch

Spitch builds speech models for African languages: text-to-speech and speech-to-text for Yoruba, Hausa, Igbo, English and Amharic, built for those languages rather than adapted from English models.

Spitch is available through the Cencori Voice API. Pick a Spitch model and every call is metered, logged and PII-redacted through the gateway, with spend caps, cost tracking and bring-your-own-key support per project.