ÒreAyò
Speech recognition preview — first working capability

Language intelligence for languages underserved by modern AI

Starting with Yorùbá, Ìgbò and Hausa.

ÒreAyò builds foundational language intelligence for the voice tools, assistants, and applications that will use it.

Speech recognition preview

Try speech recognition in Yorùbá, Ìgbò, or Hausa.

This is an early preview of one capability, not the full ÒreAyò assistant. It writes down what it hears — it does not answer questions yet.

Accuracy is still limited, and mistakes are expected — especially with background noise or fast speech. You can correct the transcript, and doing so helps us improve.

1. Choose a language

Loading languages…

2. Record or upload your speech

Audio file, up to 30 seconds.

Evaluation

Real speech examples

Recordings from people who tried ÒreAyò and agreed to share them. Each shows what the system heard and what the speaker thought of the result — including where it got things wrong.

Yorùbá

Tested with v1.0.2-asr
20 sec

ÒreAyò transcribed

orúkọ mi ni arábìrin bùsọlá inúmidùn láti jẹ ọkan lára àwọn tí wọn kópa láti gbìyànjú ọrẹ ayọ ní èdè yòrùbá ní orídì ède nàìjíríà

Tester assessment: Accurate

Show details

English translation

My name is Madam Bùsọlá. I am happy to be one of those taking part in trying ÒreAyò in the Yorùbá language in Nigeria.

Added by a person, not by ÒreAyò.

Consented public test submission

Yorùbá

Tested with v1.0.2-asr
9 sec

ÒreAyò transcribed

inú midùn láti bìyànju ọrẹ ayàn ní èdè yòrùgbá

Tester assessment: Mostly accurate

Show details

Tester's correction

inú midùn láti gbìyànju òreayò ní ède yorùbà

English translation

I am happy to try ÒreAyò in the Yorùbá language

Added by a person, not by ÒreAyò.

Consented public test submission

Yorùbá

Tested with v1.0.2-asr
5 sec

ÒreAyò transcribed

ọlọ́run dá ayé àti ọrun ayé sì wà ní jú ojò

Tester assessment: Mostly accurate

Show details

Tester's correction

ọlọ́run dá ayé àti ọrun ayé sì wà ní jùjú

English translation

God created the earth and the heavens, and the earth was without form and void.

Added by a person, not by ÒreAyò.

Consented public test submission

Ìgbò

Tested with v1.0.2-asr
9 sec

ÒreAyò transcribed

nne m kaekimere ahụ ọdịkwa ghị iriele nri taa nnọọ

Tester assessment: Accurate

Consented public test submission

Hausa

Tested with v1.0.2-asr
17 sec

ÒreAyò transcribed

ina kwana ya gajiya lafiya kalau ya jikin ya yarinya ya komai da komai

Tester assessment: Accurate

Show details

Tester's correction

none

Consented public test submission

Language should not be a barrier to technology.

Voice technology has advanced rapidly, but it still leaves millions of people behind.

Most voice assistants are built for a small set of global languages, often failing to understand local speech, accents, and everyday expressions.

ÒreAyò exists to bridge that gap by starting with the languages Nigerian families already speak, not as an add-on, but as the foundation.

What is ÒreAyò?

Built on real understanding

ÒreAyò is built to understand natural speech properly, so that the voice tools and assistants built on it can respond clearly, without requiring technical knowledge.

Built for Nigerian languages

From tone and pronunciation to code-switching, ÒreAyò is designed with Nigerian languages and speech patterns in mind — not adapted from a system built for English.

Human-centred

ÒreAyò prioritises clarity, accessibility, and ease of use, especially for parents, elders, and everyday users.

What ÒreAyò is being built to do

The capabilities below represent ÒreAyò's intended direction. The full assistant is not publicly available yet.

Conversational voice assistance

Natural, spoken conversations in indigenous languages.

Ask and understand

Ask questions in your own language and receive clear, spoken responses.

Voice-assisted actions (early stage)

Voice notes, explanations, and guided step-by-step help.

Language inclusion and accessibility

Making AI usable for parents, elders, and communities often excluded from modern technology.

Language support roadmap

Pre-release status
YorubaSpeech validation
IgboSpeech validation
HausaSpeech validation
Nigerian PidginExploration

Built responsibly from day one.

ÒreAyò is being developed with a strong focus on privacy, safety, and ethical research practices.

We prioritise transparent progress and responsible data use.

Clear communication about limitations is part of building trust especially during early research stages.

Learn about safety and privacy →

Follow the journey.

ÒreAyò is currently in active research and development, with its first speech recognition capability now in public testing and evaluation.

Join the waitlist to receive updates as the project progresses, or follow our research journey as we build toward public access.